Fact-checked by Grok 2 weeks ago

DEC Alpha

The DEC Alpha, also known as Alpha AXP, is a 64-bit reduced instruction set computing (RISC) developed by (DEC) and introduced in 1992 as the successor to the company's VAX line of complex instruction set computing (CISC) systems. It features a load/store design with 32-bit fixed-length instructions, 32 general-purpose 64-bit integer registers (with R31 reading as zero), 32 64-bit floating-point registers (with F31 reading as zero), and a flat 64-bit virtual little-endian byte-addressable memory model supporting up to 264 bytes without segmentation. The architecture was engineered for high performance and long-term scalability, targeting a 1,000-fold improvement in processing power over 25 years through support for high clock speeds, multiple instruction issue, multiprocessor configurations, and future technologies like 128-bit data paths. Development of Alpha began in as a DEC initiative to modernize the VAX and retain its customer base, evolving from the project at DEC's Systems Research Center and officially sanctioned in October 1989. Involving over 2,000 engineers across hardware and software teams, the project emphasized RISC principles for efficiency, avoiding dependency in favor of PALcode (Privileged Architecture Library code) for low-level operating system functions, and ensuring no bias toward specific programming languages or operating systems like or (DEC's UNIX variant). The first implementation, the DECchip 21064 , was a single-chip fabricated on a 0.75-micrometer process with 1.68 million transistors, operating at up to 200 MHz and achieving peak performance of 400 (million instructions per second) and 200 MFLOPS (million floating-point operations per second). This chip powered initial systems shipped in late 1992, including the DEC 3000 AXP workstations, DEC 4000 AXP departmental servers, and DEC 7000/10000 AXP midrange/mainframe platforms, which supported up to 16 processors, 14 GB of , and terabyte-scale . Alpha's defining characteristics included relaxed memory ordering with explicit barriers (memory barrier [MB] and implant bit barrier [IMB] instructions) for multiprocessor coherence, support for both IEEE and VAX floating-point formats, and binary translation tools like VEST for migrating VAX applications, enabling translated code to run at 1.05–1.7 times the speed of native VAX equivalents on early Alpha hardware. Subsequent generations, such as the 21164 (EV56) in 1994 and 21264 (EV6) in 1996, pushed clock speeds to over 600 MHz, incorporated multimedia extensions like Motion Video Instructions (MVI), and maintained leadership in benchmarks like SPECmark89, often outperforming contemporaries like MIPS R4000 or early x86 processors. The architecture's LP64 programming model—where pointers and long integers are 64 bits—facilitated large-memory systems exceeding 4 GB by 1994, influencing industry standards for 64-bit computing. Following DEC's acquisition by in 1998, Alpha development continued briefly under the new ownership, with the final major release, the EV8 (21464), planned but canceled in 2001 in favor of Intel's architecture. phased out Alpha production by 2001, though some systems remained in use for into the 2000s, and solutions now sustain legacy applications. Alpha's innovations in 64-bit RISC design and binary contributed to the broader shift toward 64-bit architectures in modern processors.

History

Origins in PRISM and RISCy VAX

In the mid-1980s, (DEC) initiated at its Western Research Laboratory in , as part of a broader effort to develop a reduced instruction set computing (RISC) architecture to succeed the aging VAX complex instruction set computing (CISC) line. The primary goals of were to simplify the instruction set for easier implementation and higher performance, incorporate deep pipelining to exploit , and support both and UNIX operating systems while maintaining with VAX software through techniques. Key figures in 's development included Richard L. Sites, who contributed to architectural explorations, and the project emphasized a 32-bit design optimized for workstations and mid-range servers. By 1988, amid growing competition from RISC architectures like and , DEC formed the RISCy VAX Task Force—also known as the Extended VAX (EVAX) group—to assess how to evolve the VAX lineage without fully abandoning its software ecosystem. This effort culminated in the 1989 RISCy VAX prototype, a 32-bit that integrated VAX compatibility modes with RISC principles such as load-store operations and a streamlined to boost performance while minimizing disruption to existing applications. The prototype, led by figures including Sites and explored at DEC's research labs, aimed to deliver incremental improvements over VAX through hardware support for and subsetted instructions. However, evaluations in late revealed limitations in the 32-bit addressing and compatibility overhead, prompting DEC to abandon RISCy VAX by early 1990 in favor of a clean-slate 64-bit RISC unencumbered by VAX legacies. This decision, formalized in fall when the Alpha project received official sanction, reflected strategic priorities for future-proof scalability and competitiveness in . Alpha thus inherited core RISC tenets from and RISCy VAX, including simplified decoding and pipelined execution, to form the basis of DEC's next-generation processor family.

Development of the Alpha Architecture

The development of the Alpha architecture by (DEC) marked a strategic pivot to a pure 64-bit reduced instruction set computing (RISC) design in the early , aimed at achieving superior and long-term to succeed the aging VAX systems. The project, initially explored through a in 1988, gained formal approval as an advanced development program in fall , with conceptual work solidifying into a comprehensive strategy by late 1989. Specifications were finalized by July 1990, transitioning to full product development that summer, reflecting DEC's commitment to a clean-slate architecture unencumbered by legacy constraints. This effort built briefly on influences from the earlier project, adopting simplified instruction principles while pursuing a fully 64-bit foundation from inception to future-proof against growing memory demands. Key design decisions emphasized performance through a load-store , which separated memory access from computation to enable efficient , and the elimination of condition codes in favor of register-based predicates for branching and conditional moves, avoiding bottlenecks in status registers. The was optimized for deep —such as the 7-stage integer pipeline in early implementations—and superscalar execution, supporting dual-issue capabilities to process multiple where feasible. These choices prioritized hardware simplicity and speed, targeting a 25-year lifespan with goals of over 100 performance and seamless migration from VAX and environments via tools. The first implementation, known as EV4 or DECchip 21064, saw its processor module power up in June , followed by a full system in September and successful booting of on September 9, . First silicon for the 21064 arrived in late , with public announcement of the Alpha architecture occurring in February 1992 and volume shipment beginning in September 1992. This timeline underscored DEC's aggressive push, culminating in the November 1992 debut of Alpha-based systems like the DEC 3000 series. Unique to Alpha's formal definition were its 64-bit flat —initially utilizing 43 bits but architecturally extensible to full 64 bits—a little-endian byte order for data alignment, and the deliberate exclusion of in favor of PALcode (Privileged Architecture Library code) for operating system-specific operations, enhancing execution speed and flexibility across environments. These specifications ensured a scalable, high-performance foundation, positioning Alpha as a competitive 64-bit RISC platform for workstations and servers.

Evolution of Alpha Models

The evolution of Alpha models began with the introduction of the EV4 (DECchip 21064) in 1992, operating at up to 200 MHz and marking Digital Equipment Corporation's (DEC) entry into high-performance 64-bit RISC computing. This initial implementation featured on-chip primary caches of 8 KB each for instructions and data, with external secondary caching, and was fabricated using a 0.75 μm process to achieve rapid clock speeds competitive with contemporary ECL-based systems. The EV4 established the foundation for Alpha's hardware lineage, emphasizing simplicity and speed in its dual-issue pipeline design. In 1995, DEC advanced the architecture with the EV5 (DECchip 21164), also known as the LCA in some low-cost variants, reaching 300 MHz and introducing significant integration improvements, including a 96 on-chip second-level cache shared between instruction and data streams. This model shifted to a quad-issue superscalar , enhancing throughput while maintaining compatibility with the original Alpha (ISA), which remained unchanged across generations. Fabricated on a 0.50 μm process, the EV5 also supported variants like the EV56, which widened the external bus to 128 bits for better in systems requiring higher data transfer rates. The EV6 (DECchip 21264), released in 1998 at initial speeds of 450-500 MHz with later variants reaching 600 MHz, represented a major leap in clock speed and , incorporating and a dedicated EV6 bus protocol that enabled point-to-point connections with transfers providing peak bandwidths up to approximately 6.4 GB/s in later implementations. Built on a 0.35 μm process, it integrated larger on-chip L1 caches (64 KB instruction and 64 KB data) while relying on external caching, prioritizing for multiprocessor configurations. This generation solidified Alpha's position in high-end workstations and servers, with the EV6 bus later licensed to other vendors for broader ecosystem compatibility. Following Compaq's acquisition of DEC in 1998 and the subsequent merger with in 2002, the EV7 represented the final major Alpha microprocessor. By 2002, the EV7 (DECchip 21364), codenamed , introduced directory-based to support scalable multiprocessor systems, featuring integrated L2 cache up to 1.75 MB and point-to-point interconnects for up to 128 processors. Operating at speeds around 1 GHz on a 0.18 μm process, it emphasized low-latency networking and error-correcting code ( protection, targeting enterprise and scientific computing applications, with production continuing until systems like the ES80 and GS1280 were discontinued around 2006-2007. The planned EV8, an ambitious 8-wide superscalar design aiming for even higher issue rates and integration, was canceled in June 2001 amid shifting priorities. In 1994, DEC simplified branding by dropping "AXP" from the Alpha name, reflecting its maturation as a standalone architecture. However, DEC's acquisition by in 1998 accelerated the platform's decline, as Compaq prioritized Intel's ecosystem, announcing Alpha's phase-out by 2004 while completing existing commitments.

Architectural Design

Core Design Principles

The DEC Alpha architecture embodies core RISC (Reduced Instruction Set Computing) principles, prioritizing simplicity and efficiency to achieve high performance through streamlined instruction execution. It employs fixed-length 32-bit instructions, all aligned on longword boundaries, which facilitates uniform decoding and enables effective pipelining across implementations. The design adheres strictly to a load-store model, where memory operations are isolated from computational instructions, requiring data to be loaded into registers for arithmetic and logical processing, thereby minimizing memory access latency and supporting parallel execution. This register-rich approach, with dedicated sets for integer and floating-point operations, further reduces reliance on memory, allowing multiple instructions to proceed concurrently without data dependencies hindering throughput. A key tenet is the avoidance of condition codes, which eliminates hidden state updates that could complicate pipelining and multiple instruction issue. Instead, comparisons produce results directly in registers, often comparing against for branching decisions, enhancing predictability and efficiency in . Branch prediction is integrated via static rules—forward branches predicted as not taken and backward as taken—supplemented by optional hints to guide dynamic hardware predictors, without the use of delay slots that might impose software overhead. The architecture designates register R31 (and F31 for floating-point) as implicitly , hardwired to read as zero while ignoring writes, which simplifies zero-extension operations and immediate value handling in comparisons. Pipelining forms a foundational , with the assuming deep pipelines to overlap fetch, decode, execute, and commit stages, as seen in early models featuring 7-stage and 10-stage floating-point pipelines. is encouraged through mechanisms like conditional moves and branch hints, permitting where dependent operations can proceed provisionally, with precise via trap barriers to maintain correctness. This approach tolerates reordering of loads and stores within the same processor, using memory barriers only when strict ordering is required, to maximize without architectural penalties. From its inception, Alpha was engineered as a pure 64-bit architecture, eschewing any 32-bit to enable seamless handling of large address spaces—up to 2^64 bytes virtually—and native 64-bit integer operations on quadword . This design choice, rooted in Digital's project exploration of RISC concepts, ensures scalability for future performance demands without legacy constraints.

Registers and Addressing

The DEC Alpha architecture features a large register file consisting of 31 general-purpose 64-bit integer registers, designated R0 through R30, with R31 serving as a hardwired zero register that always reads as zero and discards writes to it. This design provides ample registers for computations in its load-store paradigm, where R0 through R30 can be used freely by software, while the zero register simplifies comparisons and certain operations by eliminating the need for explicit zero-initialization instructions. Complementing the integer registers are 31 64-bit floating-point registers, F0 through F30, with F31 also functioning as a zero register under the same rules. Alpha employs a simplified set of addressing modes aligned with its RISC principles, primarily register-indirect with a signed for load and operations, computed as the base register value plus a 16-bit sign-extended . There is no support for immediate addressing in arithmetic operations beyond these displacements, nor complex modes like scaled indexing in memory references; however, scaled multiply-add instructions (e.g., S8ADD for octet ) allow software to emulate indexed addressing for accesses. Branch instructions utilize PC-relative addressing with a 21-bit signed , enabling jumps within a ±1 MB range relative to the for efficient . Software conventions define specific roles for certain registers to support procedure calls and stack management. R30 serves as the stack pointer (SP), pointing to the top of the current stack and growing downward, while R15 acts as the frame pointer (FP) to delineate stack frames. The return address for procedure calls is stored in a designated register , conventionally R26, by jump and branch instructions like JSR and BSR. Stack frames are typically aligned to 16 bytes, with additional per-operating-system variations for handling and stacks. The Alpha employs a flat 64-bit virtual addressing model without segmentation, where all addresses are treated uniformly in a linear space. Virtual-to-physical translation uses a multi-level structure or translation buffer, with a minimum supported of 43 bits (8 terabytes) but architecturally capable of the full 64 bits (16 exabytes); implementations varied in size due to constraints. Page sizes are implementation-dependent but default to 8 , with support for larger superpages up to 64 to reduce translation overhead.

Data Types and Memory Model

The DEC Alpha architecture supports four primary integer data types: byte (8 bits), word (16 bits), longword (32 bits), and quadword (64 bits), all represented in format for signed values, with unsigned interpretations available through arithmetic instructions that preserve the bit pattern. However, the base architecture provides native operations only for longword and quadword integers, with loads zero-extending smaller types where supported; byte and word operations, including , require multi-instruction sequences or extensions like BWX for dedicated instructions such as SEXTB and SEXTW. Sign extension for loaded values is typically achieved using compare and shift operations. For floating-point data, the architecture adheres to the standard with single-precision (S_floating, 32 bits) and double-precision (T_floating, 64 bits) formats, including support for denormalized numbers, infinities, (both signaling and quiet), and configurable rounding modes (normal, chopped, plus/minus infinity). Additionally, it accommodates VAX legacy formats like F_floating (32 bits) and G_floating (64 bits) for compatibility, with conversions via CVT instructions. (extended precision) uses the 128-bit X_floating format, implemented in software and spanning two adjacent 64-bit floating-point registers for storage and operations. The memory model employs , also known as relaxed consistency, permitting compiler and hardware reordering of memory operations unless constrained by explicit synchronization primitives like instructions (MB for all barriers, WMB for write barriers) or load-locked/store-conditional pairs (e.g., LDQ_L/STQ_C) to enforce and atomicity. is maintained through implementation-specific protocols, such as variants of MESI in multiprocessor systems, with the IMB instruction ensuring instruction consistency across processors. Alignment requirements mandate natural boundaries—1 byte for bytes, 4 bytes for longwords and single-precision floats, and 8 bytes for quadwords and double-precision floats—to avoid exceptions, though octaword (16-byte) alignment is recommended for optimal performance in paired operations. Alpha systems operate in little-endian byte order by default, where the least significant byte is stored at the lowest address, with optional big-endian support configurable at time via address (e.g., inverting VA<2> for longword accesses). Unaligned memory accesses are permitted without mandatory traps, handled transparently by instructions like LDQ_U and STQ_U with a performance penalty, but implementations may optionally generate alignment faults for correction in software.

Instruction Set

Instruction Formats and Encoding

The DEC Alpha employs a fixed-length instruction set where all instructions are 32 bits wide and must be aligned on 4-byte boundaries. This design adheres to RISC principles by using uniform 32-bit encodings without variable-length instructions, ensuring straightforward decoding. The instructions are divided into three primary formats—, , and Operate—each optimized for specific operations while sharing a common 6-bit primary field in bits 31:26. The Branch format supports control-flow instructions and consists of the 6-bit , a 5-bit source register specifier (Ra, bits 25:21), and a 21-bit signed (bits 20:0). The is sign-extended and shifted left by 2 bits to form a byte offset from the current , providing a branch range of approximately ±1 million instructions. Opcodes in this format, such as 30₁₆ for unconditional branches, allocate specific values within the primary opcode space. The Memory format is used for load and store operations, featuring the 6-bit and a 16-bit signed (bits 15:0). For loads, bits 25:21 specify the destination () and bits 20:16 the base (), with the effective computed as R[Rb] + sign-extended . For stores, bits 25:21 specify the base () and bits 20:16 the source (), with the effective computed as R[Ra] + sign-extended . Rb serves as the base for loads (not a scaled index) and as the source for stores. Example opcodes include 08₁₆ for load instructions. The Operate format handles arithmetic, logical, and other computational instructions, with the 6-bit opcode, three 5-bit register specifiers (Ra, Rb, and Rc in bits 25:21, 20:16, and 15:11 respectively for sources and destination), and a 6-bit function field (bits 10:5) for both integer and floating-point operations. The Memory format supports up to 16-bit signed immediates for address calculations. Integer opcodes like 10₁₆ and floating-point subsets such as 15₁₆ or 16₁₆ fall under this format. Opcode allocation uses the 6-bit primary to categorize instructions, with dedicated subsets for floating-point operations (e.g., opcodes 10₁₆, 15₁₆) and privileged PALcode instructions ( 00₁₆, using the function for sub-operations like system calls). specifiers (Ra, Rb, Rc) are uniformly 5 bits each, addressing the 32 general-purpose or 32 floating-point registers (values 0–31, where 31 often denotes zero). Unused specifier default to 31.
FormatBits 31:26 (Opcode)Bits 25:21 (Ra)Bits 20:16 (Rb)Bits 15:0 (Displacement/Function/Literal)Key Features
Branch6-bit primary5-bit Higher bits of 21-bit signed (bits 20:0, with low 16 bits in 15:0)Sign-extended, shifted left by 2 for offset
Memory6-bit primaryDestination (loads) / Base (stores)Base (loads) / Source (stores)16-bit signed For loads: = R[Rb] + sext(disp); for stores: = R[Ra] + sext(disp)
Operate6-bit primary5-bit 5-bit 6-bit field (bits 10:5) for and floating-point operations; Rc in bits 15:11Register-register operations; no immediates in this format

Load-Store Operations

The DEC Alpha architecture employs a load-store design, where data movement between the register file and memory is exclusively handled by dedicated load and store instructions, ensuring a clean separation from computational operations. This approach supports efficient pipelining and in implementations. All memory accesses use virtual addressing, with the memory model assuming little-endian byte ordering for multi-byte types. Integer load instructions transfer data from to the 64-bit registers (R0–R31), with specific variants for different sizes and extension behaviors. The LDL instruction loads a 32-bit signed longword from , sign-extending it to 64 bits in the destination , and requires 4-byte . For 64-bit quadwords, LDQ loads the full value without extension, aligned on an 8-byte boundary. Unsigned smaller loads include LDBU for 8-bit bytes (zero-extended to 64 bits) and LDWU for 16-bit words (zero-extended), both without restrictions. A typical syntax is LDL R1, offset(R2), which loads a signed longword from the effective R2 + offset into R1. Floating-point loads move data from to the floating-point registers (F0–F31), supporting both VAX and IEEE formats with corresponding levels. LDF loads a 32-bit VAX F_floating single- value, aligned on 4 bytes, while LDG loads a 64-bit VAX G_floating double- value, aligned on 8 bytes. For IEEE , LDS handles single-precision S_floating and LDT double-precision T_floating, also with standard alignment. If the destination floating-point register specifier (Fa) is 31, these instructions (LDF, LDG, LDS, LDT) function as prefetches rather than loads, bringing data into the without altering registers. Integer store instructions reverse the process, writing from integer registers to memory while enforcing size and alignment rules. STL stores the low 32 bits of a register as a signed longword (4-byte aligned), and STQ stores the full 64-bit quadword (8-byte aligned). Smaller stores like (8-bit byte) and (16-bit word) have no alignment requirements. Syntax follows a similar pattern, such as STQ R1, offset(R2), which stores the quadword from R1 to the effective address. Floating-point stores transfer from floating-point registers to memory, mirroring the load formats. STF stores a 32-bit F_floating value (4-byte aligned), and STG a 64-bit G_floating value (8-byte aligned). IEEE stores use for S_floating and STT for T_floating, with equivalent alignment. To handle unaligned accesses, Alpha provides specialized instructions like LDQ_U for unaligned quadword loads and STQ_U for unaligned stores, which do not require 8-byte boundaries but may incur performance penalties in hardware implementations. Unaligned accesses generally trigger a data alignment exception (offset 280₁₆ in the system control block), handled by privileged architecture library (PALcode) routines such as ealnfix for even-odd alignment fixes or dalnfix for dynamic handling, saving the faulting virtual address in R4 and operation type (0 for read, 1 for write) in R5. Prefetch operations prepare data for future loads without immediate register updates, using the PREFETCH instruction in variants like PREFETCH_M (for modification intent, loading to level-1 cache in modified state) or PREFETCH_EN (evicting the next line). Probe instructions, implemented via PALcode calls (PROBER for read access and PROBEW for write), check memory accessibility and permissions without performing the access, returning success or exception details to support memory management.
CategoryKey InstructionsPurpose and Notes
Integer LoadsLDL, LDQ, LDBU, LDWUSize-specific transfers with sign/zero extension; alignment enforced except for byte/word.
Floating-Point LoadsLDF, LDG, , LDTVAX/IEEE formats; all prefetch if Fa=31.
Integer StoresSTL, STQ, STB, STWReverse of loads; no extension needed.
Floating-Point StoresSTF, STG, , STTFormat-preserving writes.
Unaligned/SpecialLDQ_U, STQ_U, PREFETCH, PROBER/PROBEWHandle misalignment, caching, and access checks; exceptions via PALcode.

Arithmetic and Logical Instructions

The DEC Alpha architecture provides a comprehensive set of register-to-register arithmetic and logical instructions for both integer and floating-point operations, emphasizing simplicity and performance through a load-store design. These instructions operate on the 32 general-purpose integer registers (R0–R31, with R31 always reading as zero) and 32 floating-point registers (F0–F31, with F31 always reading as zero), enabling efficient data manipulation without direct memory access. Overflow detection and exception handling are integrated to support robust computation, with traps configurable via instruction qualifiers or control registers. Integer arithmetic instructions include (ADD), (SUB), and (MUL for the low 64 bits, UMULH for the high 64 bits of unsigned multiply). These are available in variants for 64-bit (), 32-bit (), and unqualified forms that treat operands as 64-bit. For example, the ADD instruction computes Rc = Ra + Rb, writing the least significant 64 bits of the result, while qualifiers like /V (overflow trap) or /S (software-complete trap) enable arithmetic overflow detection by signaling an (IOV) trap if the result exceeds the representable range. Integer division is not supported in and must be emulated in software, often using multiply-based algorithms. These operations prioritize speed, with no default traps to avoid performance penalties in non-critical code paths. Logical operations encompass bitwise AND (AND), OR (OR, also known as BIS), and exclusive-OR (XOR), each performing the respective operation on 64-bit operands and storing the result in Rc without any traps or side effects. The conditional move (CMOV) instruction enhances logical efficiency by using predicates—such as equality (CMOVEQ), less than (CMOVLT), or zero/non-zero conditions on —to selectively copy Rb to Rc, avoiding branches and enabling predicate-based optimization in compilers. For instance, CMOVEQ Ra, Rb, Rc moves Rb to Rc only if Ra is zero, supporting compact implementations of constructs directly in hardware. These instructions operate uniformly on all 64 bits, facilitating for flags, masks, and data packing. Floating-point arithmetic supports IEEE 754 single-precision (S_floating, F_floating) and double-precision (T_floating, G_floating) formats through instructions like addition (ADDF for single, ADDG for double), subtraction (SUBF, SUBG), multiplication (MULF, MULG), and division (DIVS, DIVT, DIVF, DIVG). For example, ADDF Fa, Fb, Fc adds the single-precision values in Fa and Fb, rounding the result to Fc according to modes specified by qualifiers (/C for chopped, /M for minus infinity, /D for plus infinity) or the Floating-Point Control Register (FPCR), which also governs dynamic rounding to nearest or toward zero. These operations set exception flags in the FPCR for inexact (INE), underflow (UNF), overflow (OVF), division by zero (DZE), and invalid operation (INV), with traps enabled via the FPCR's trap enable bits; using F31 as the destination may suppress traps for non-trapping computations. Rounding modes ensure compliance with IEEE standards while supporting legacy VAX behaviors, balancing precision and performance in scientific applications. Shift and extract instructions extend logical capabilities for data alignment and manipulation, including logical left shift (SLL), arithmetic right shift (), and extract word low (EXTWL). SLL shifts Ra left by the low 6 bits of Rb (0–63 positions), filling with zeros and discarding high bits into Rc, while performs a signed right shift, preserving the through extension. EXTWL extracts a byte or word from Ra at an offset specified by the low 3 bits of Rb (multiples of 8 bytes), sign-extending it to a 64-bit value in Rc for efficient unaligned handling. These instructions, operating solely on registers, integrate seamlessly with operations to support variable-length data processing without traps.
Instruction CategoryKey ExamplesNotable Features
Integer ArithmeticADD, SUB, MUL, UMULHOverflow traps via /V or /S; 64-bit results by default; division emulated in software.
Logical OperationsAND, OR, XOR, CMOVBitwise on 64 bits; CMOV uses predicates for branch-free code.
ADDF/SUBG/MULF/DIVSIEEE rounding modes; FPCR-managed exceptions (INE, OVF, etc.).
Shifts and ExtractsSLL, , EXTWL0–63 bit shifts; in SRA and EXTWL.

Control and Branch Instructions

The control and branch instructions in the DEC Alpha architecture manage program flow by altering the (PC) and handling subroutine calls, conditional execution, and exceptions, adhering to the principle of avoiding dedicated condition codes by directly testing registers for decisions. Unconditional branches include the BR instruction, which performs a PC-relative jump to a target address without saving a , and the BSR (branch to subroutine) instruction, which executes a similar PC-relative while storing the address of the following in register 26 (the return address register, RA). The BSR supports subroutine calls within a displacement range of approximately ±1 MB (a signed 21-bit field scaled by 4 bytes for longword ). These instructions facilitate straightforward jumps and invocations, with BSR commonly used for short-range calls in compiled code. Conditional branches in Alpha test or floating-point registers directly rather than flags. For , instructions like BEQ (branch if equal) and BNE (branch if not equal) compare the source register Ra to zero, branching on or across all 64 bits treated as a signed quadword; these use a PC-relative displacement of up to ±1 million instructions. Additional conditionals, such as BLBC (branch if low bit clear) and BLBS (branch if low bit set), examine the least significant bit of Ra for bit-level decisions. Floating-point branches, exemplified by FBEQ (floating-point branch if equal), test the source floating-point register for to zero (considering and exponent), with complementary forms like FBNE for ; these also employ the ±1 million instruction range and support T_floating format comparisons. Such designs enable efficient predicate evaluation without intermediate condition . Register-based jumps handle longer-range or computed control transfers, including JMP for unconditional jumps to an address in Rb (often used for table-driven loops), JSR (jump to subroutine) which jumps to Rb while saving the return address in RA, and RET (return) which jumps using the address in RA to exit procedures. These operate over the full 64-bit virtual address space and include hint bits to aid branch prediction, such as distinguishing calls from returns. Loops are typically implemented using conditional branches combined with these jumps for iteration control. Exception handling integrates with control flow via the CALL_PAL instruction, which invokes privileged architecture library (PALcode) routines for operating system calls, interrupt returns, or hardware management, clearing the lock flag and stalling prior instructions for serialization. Hardware traps, such as those for arithmetic errors (e.g., integer overflow or floating-point underflow) or unaligned accesses, are signaled asynchronously and vectored through PALcode, with the processor advancing the PC past the trapping instruction and saving the faulting or next address for handler use. Return address management ensures reliable subroutine and exception recovery by preserving the PC in RA or stack frames during traps. Barriers like TRAPB (trap barrier) prevent speculative execution across potential arithmetic traps, maintaining precise exception semantics.

Extensions and Variants

Byte-Word Extensions (BWX)

The Byte-Word Extensions (BWX) were an optional addition to the DEC Alpha architecture, first implemented in the Alpha 21164A (EV56) in , to support efficient handling of sub-word data types without requiring software or complex instruction sequences. This extension addressed the original Alpha design's focus on 32-bit (longword) and 64-bit (quadword) operations by introducing hardware support for 8-bit (byte) and 16-bit (word) memory accesses, enhancing performance in environments like Unix where byte-level manipulations are common. BWX added four primary memory access instructions: LDBU for loading an unsigned byte into the low 8 bits of a register (zero-extending the rest), LDWU for loading an unsigned word into the low 16 bits (zero-extending the rest), STB for storing a byte from the low 8 bits of a register, and STW for storing a word from the low 16 bits. BWX also added instructions for extracting, inserting, masking, and zeroing bits within quadwords, such as EXTBL, INSBL, MSKBL, ZAP, and ZAPNOT. The store instructions incorporate address-based masking, using the two low-order bits of the virtual address to select which bytes or bits within the target quadword are written, thereby enabling partial updates without affecting adjacent data. Complementary register-to-register instructions like SEXTB (sign-extend byte) and SEXTW (sign-extend word) were also included to handle sign extension efficiently, reducing overhead in operations involving signed sub-word data. As an optional extension, BWX presence is detected at runtime using the AMASK instruction, which clears bit 0 in the result if supported, allowing operating systems and applications to probe hardware capabilities dynamically. This approach ensures with earlier Alpha processors like the 21064A, where unsupported BWX instructions would trap and emulate via software, though at a significant performance cost. By providing native byte and word operations, BWX improved portability of programs and software to Alpha without resorting to a 32-bit , as it allowed direct manipulation of heterogeneous data structures common in these environments. The extension's impact was particularly notable in string processing and I/O tasks, where byte-level accesses previously incurred significant performance penalties due to unaligned loads or multi-instruction emulation; BWX reduced this overhead by enabling aligned, granular memory operations and minimizing sign-extension sequences. Overall, it elevated Alpha's suitability for general-purpose by bridging the gap between its 64-bit RISC foundation and legacy byte-oriented codebases.

Multimedia and Specialized Extensions (MVI, FIX, CIX)

The and Specialized Extensions in the DEC Alpha encompassed three optional set extensions—MVI, FIX, and CIX—designed to enhance performance in specific domains without introducing full vector processing units. These extensions were implemented in later Alpha generations, providing targeted accelerations for processing, floating-point operations, and tasks. They were detected via the AMASK , which returned specific bits to indicate hardware support; absent support triggered illegal traps for software . The Motion Video Instructions (MVI) extension, introduced in 1997 with the Alpha 21164PC (PCA56) microprocessor and also supported in the Alpha 21264 (EV6), added 13 single-instruction multiple-data (SIMD)-like operations operating on packed bytes and words within the 64-bit integer registers to accelerate image and video processing algorithms. These instructions supported tasks such as pixel value comparisons for motion estimation, data packing/unpacking for format conversions, and error calculations in video decoding pipelines like MPEG-1 and MPEG-2. Unlike broader SIMD extensions in other architectures, MVI focused on unsigned and signed saturated arithmetic to prevent overflow in multimedia computations, with a typical latency of 2-3 cycles on supported hardware. Key instructions included:
  • Byte and word minimum/maximum operations: MINUB8 (minimum unsigned bytes), MAXUB8 (maximum unsigned bytes), MINSB8 (minimum signed bytes), MAXSB8 (maximum signed bytes), MINUW4 (minimum unsigned words), MAXUW4 (maximum unsigned words), MINSW4 (minimum signed words), and MAXSW4 (maximum signed words), used for clamping and comparing intensities.
  • Packing and unpacking: PKWB (pack words to bytes), UNPKWB (unpack bytes to words), PKLB (pack longs to bytes), and UNPKBL (unpack bytes to longs), facilitating and expansion of video data.
  • Pixel error: PERR (sum of absolute byte differences), essential for block-matching in during video .
This extension optimized real-time video authoring and playback without requiring dedicated vector registers, integrating seamlessly with the base Alpha integer pipeline. The Floating-Point Extensions (FIX), introduced in 1998 with the (EV6) microprocessor, augmented the base with nine additional instructions to improve compliance and accelerate common operations like square roots and format conversions between integers and floating-point values. These enhancements addressed limitations in earlier Alpha implementations by supporting efficient transfers between VAX (F_floating single-precision, G_floating double-precision) and IEEE (S_floating single-precision, T_floating double-precision) formats, as well as direct integer-to-floating-point moves to reduce software overhead in numerical computations. The instructions utilized the existing 32 floating-point registers and Floating-Point Control Register (FPCR) for rounding and , maintaining the architecture's emphasis on high-performance pipelined execution. Representative instructions included:
  • Conversion operations: FTOIS (S_floating to signed integer), FTOIT (T_floating to signed integer), ITOFF (integer to F_floating), ITOFS (integer to S_floating), and ITOFT (integer to T_floating), enabling fast transitions in mixed-precision algorithms.
  • Square root instructions: SQRTF (F_floating ), SQRTG (G_floating ), SQRTS (S_floating ), and SQRTT (T_floating ), providing hardware acceleration for iterative methods in scientific and graphics applications.
FIX was particularly valuable for workloads requiring precise , such as simulations, where it reduced compared to emulated operations on pre-EV6 processors. The Count Extensions (CIX), introduced in 1999 with the Alpha 21264A (EV67) microprocessor, provided three instructions for efficient bit counting and scanning on 64-bit operands, targeting algorithms in data , hashing, and that rely on population counts and zero positioning. These operations filled a gap in the base Alpha set by enabling hardware-level support for tasks like in or parity checks in error detection, without the need for loop-unrolled software routines. CIX instructions executed in the pipeline, with low to support frequent use in string processing and manipulations. The instructions were:
  • CTLZ (count leading zeros), which returns the number of leading zero bits in a 64-bit , useful for in floating-point or alignment in hashing.
  • CTPOP (count population), which tallies the number of set bits (ones) across the , applied in -based encoding for and .
  • CTTZ (count trailing zeros), which counts trailing zero bits, aiding in bit scanning for sparse structures and division optimizations.
As an optional extension, CIX enhanced the Alpha's utility in data-intensive applications by providing these primitives natively, improving throughput over software implementations on earlier cores.

Implementations

Microprocessor Generations

The DEC Alpha evolved through several s, each introducing architectural improvements to enhance while maintaining the core 64-bit RISC . The first , known as EV4 or the 21064, debuted in 1992 as a dual-issue with a separate (FPU) for handling floating-point operations. It contained 1.68 million transistors and operated at clock speeds up to 200 MHz on a 0.75 μm process, enabling high for its era through pipelined execution and branch prediction. The second generation, EV5 or 21164, arrived in and marked a shift to integrated on-chip to reduce , including 8 and 8 data L1 alongside a 96 unified . The core logic featured approximately 2.8 million transistors, with total die count reaching 9.3 million including , and supported clock frequencies from 300 MHz to 633 MHz via a 128-bit Alpha EV5 for improved . This generation emphasized superscalar issue of up to four , enhancing integer and floating-point throughput without . Succeeding it, the EV6 or 21264 in 1998 introduced to tolerate latency, featuring a 20-entry issue and a 15-entry floating-point issue for dynamic scheduling. With 15 million transistors on a 0.35 μm , it achieved clock speeds of 600 MHz to 1.25 GHz, supporting and a peak issue rate of six (four , two floating-point). This significantly boosted single-threaded through deeper pipelining and larger on-chip structures. Later generations included the EV7 or 21364, released in 2000, which integrated directory-based and a quad-issue for up to four instructions per cycle, operating at 1.15 GHz to 1.65 GHz on a 0.18 μm process with 152 million transistors (including extensive on-die ). The planned EV8 or 21464, however, was canceled in 2001 amid Compaq's shift to ; it was designed for 2 GHz operation on a 0.13 μm process and would have incorporated () with 4-way support alongside an 8-wide superscalar core to improve throughput on multiprogrammed workloads.
GenerationModelYearTransistors (millions)Clock Speed (MHz)Key FeaturesProcess (μm)
EV42106419921.68Up to 200Separate FPU, dual-issue0.75
EV52116419952.8 (core) / 9.3 (total)300–633Integrated 8 KB I/D L1 caches, 128-bit bus0.5
EV621264199815600–1250Out-of-order, 20/15-entry issue queues0.35
EV7213642000152 (total)1150–1650Quad-issue, integrated 0.18
EV821464Canceled (2001)~250 (est.)~20008-wide issue, 4-way 0.13
Process technology advanced progressively from 0.75 μm in EV4 to 0.18 μm in EV7, enabling higher densities, faster clocks, and lower power through shrinks that reduced feature sizes while maintaining fabrication for compatibility and cost efficiency.

Integrated System Implementations

The Alpha 21364 , codenamed EV7, represented a significant advancement in integrated by combining the Alpha EV68 core with on-chip logic, including a directory-based controller, , and point-to-point interconnect fabric, enabling scalable cache-coherent (CC-NUMA) multiprocessor configurations. This integration allowed the EV7 to support direct processor-to-processor communication via four bidirectional links, each providing 6.4 GB/s of (32 data bits plus at 800 Mb/s per direction), with a low of 18 ns for remote accesses in small configurations. Operating at speeds up to 1.15 GHz, the EV7 facilitated the construction of large-scale systems without external switches in basic topologies, using a switchless mesh architecture that formed or shuffle interconnect patterns for fault-tolerant scaling. In the AlphaServer ES80 and GS1280 systems, the EV7 was deployed in custom server modules optimized for and enterprise workloads. The ES80 model supported up to 8 processors within a modular quad-building-block (QBB) design that integrated four EV7 CPUs per QBB, along with and I/O ports connected via a hierarchical switch fabric offering aggregate bandwidths of up to 25.6 GB/s. The GS1280 scaled to 64 processors by interconnecting multiple QBBs or drawers in a ring topology, achieving up to 51.2 GB/s aggregate interconnect while maintaining across distributed directories for up to 512 GB of per system. These implementations emphasized point-to-point interconnects over bus-based designs, reducing contention and enabling incremental expansion in rack-mounted cabinets with redundant power and cooling. Licensing of the Alpha architecture to third parties extended its potential into custom integrated systems, though adoption remained limited. In 1996, granted a worldwide to manufacture and market Alpha processors, allowing Samsung to produce variants such as the 21164 at test volumes starting that year and a 600 MHz version of the 21264 by late 1998 using advanced eight-inch fabrication. Samsung's included to future Alpha iterations, positioning it for potential applications like devices or appliances, but production focused primarily on standard forms with minimal diversification into SoCs or specialized . variants of Alpha cores in processors were explored in 1990s prototypes by DEC and partners, but saw limited commercial adoption due to the architecture's focus on high-end computing rather than low-power markets.

Applications and Impact

Alpha-Based Computing Systems

The AlphaStation series represented Digital Equipment Corporation's (DEC) primary line of Alpha-based workstations introduced in 1994, targeting high-performance computing tasks such as engineering and scientific applications. The AlphaStation 200, equipped with the EV4 (Alpha 21064) processor at speeds up to 166 MHz, featured a compact desktop form factor with support for up to 128 MB of RAM and integrated PCI/ISA buses for expansion. Similarly, the AlphaStation 250, also launched in 1994 but utilizing an upgraded EV4 variant, offered enhanced clock speeds around 200 MHz and was positioned as a mid-range option for professional users requiring 64-bit processing capabilities. These systems competed directly with contemporary RISC-based workstations from Sun Microsystems (SPARC) and Hewlett-Packard (HP-PA), providing superior integer performance in benchmarks relevant to the era's technical workloads. Complementing the AlphaStation lineup, DEC's Multia (also known as the Universal Desktop Box) served as an all-in-one, low-cost workstation released in November 1994, integrating the Alpha 21066 processor at 166 MHz into a compact, laptop-like chassis with built-in peripherals including a 2.5-inch hard drive bay, PCMCIA slots, and multimedia support. Designed for entry-level network computing and Windows NT deployment, the Multia emphasized affordability and space efficiency, with up to 64 MB of RAM and optional Ethernet connectivity, making it suitable for small office or educational environments. Despite its innovative form factor, the Multia achieved limited market success due to performance constraints compared to higher-end AlphaStations. On the server side, the series extended Alpha architecture to enterprise environments, with models like the and 4100 introduced in the mid- using the EV5 () processor. The single-processor targeted departmental use with up to 1 GB of and PCI/EISA I/O, while the supported up to four CPUs in a scalable pedestal or cabinet configuration, accommodating up to 8 GB of memory for multi-user applications. The later DS series, including the DS10 and DS20 models from the late , focused on dense, rack-mountable designs optimized for clustering, featuring dual-processor configurations with up to 8 GB of SDRAM and high-speed interconnects like Memory Channel for fault-tolerant environments. These servers enabled setups, distinguishing them from workstation-oriented systems. Alpha-based systems primarily ran DEC's proprietary operating systems, including Digital UNIX (later rebranded Tru64 UNIX), which provided a 64-bit UNIX environment with advanced clustering via TruCluster, and , offering robust multitasking and real-time capabilities. Tru64 UNIX support extended until 2012 under (HP) stewardship following DEC's acquisition by in 1998. OpenVMS remains viable on Alpha hardware through emulation solutions, with official patches available into the 2010s. Additionally, distributions supported Alpha until the early 2010s, with kernel maintenance tapering off around 2012, though community efforts persisted for legacy compatibility.

Performance Benchmarks

The DEC Alpha processors demonstrated strong performance in standardized benchmarks, particularly in floating-point workloads, across their generational . The initial EV4 (Alpha 21064) implementation, operating at 200 MHz in systems like the DEC 10000, achieved 106.5 SPECint92 and 200.4 SPECfp92, establishing an early lead over contemporaries such as the Pentium at 66 MHz, which scored around 25 SPECint92. Later models scaled these metrics significantly; for instance, the EV5 (Alpha 21164) at 300 MHz in the Alpha XL workstation delivered 7.3 SPECint95 and 9.8 SPECfp95 base scores. By the EV6 generation (Alpha 21264), performance advanced further, with an 833 MHz single-processor configuration in the API UP2000 attaining a SPECfp_base2000 score of 571, reflecting optimizations in superscalar execution. Floating-point capabilities were a hallmark of the Alpha architecture, benefiting from dedicated hardware units that enabled high throughput in scientific computing. The EV5 achieved a peak floating-point performance of 0.6 GFLOPS through its dual floating-point pipelines (one add and one multiply per cycle), each capable of executing add or multiply operations independently at clock rates up to 300 MHz. This scaled in subsequent generations; the EV7 (Alpha 21364) reached over 4 GFLOPS peak via enhanced and multiple execution units at 1.25 GHz, supporting demanding applications like dense linear algebra solvers. In , which measure sustained double-precision performance, an EV67 (a late EV6 variant) at 500 MHz delivered approximately 637 MFLOPS on n=1000 matrices, approaching 77% of its theoretical peak under optimized conditions. Key architectural factors contributed to these results, including progressive increases in clock speeds—from 200 MHz in the EV4 to 1.25 GHz in the EV7—and larger on-chip caches, such as the 64 KB split L1 in the EV6 paired with up to 2 MB off-chip L2. (IPC) also improved, with early models like the EV4 achieving around 1 IPC in integer workloads, while the out-of-order EV6 sustained 2-3 IPC on average through its quad-issue capability and 20-entry integer reorder buffer. These elements enabled efficient handling of and branch prediction in mixed workloads. Relative to competitors, Alpha processors held a notable in floating-point-intensive tasks during the mid-1990s. For example, a 400 MHz EV5 configuration delivered SPECfp95 performance comparable to a 200 MHz (~17 vs. ~19), but the showed higher efficiency per clock in floating-point workloads. This edge persisted against x86 contemporaries like the , where Alpha systems often doubled FP metrics in SPEC suites at equivalent clock rates.
GenerationClock (MHz)Example SPECint (Base)Example SPECfp (Base)L2 CachePeak FP (GFLOPS)
EV4 (21064)200106.5 (SPEC92)200.4 (SPEC92)4 MB0.4
EV5 (21164)3007.3 (SPEC95)9.8 (SPEC95)1-8 MB0.6
EV6 (21264)500311 (SPEC2000)571 (SPEC2000 at 833 MHz)2 MB1.0
EV7 (21364)1250~400 (SPEC2000 est.)>1000 (SPEC2000 est.)4-16 MB>4.0

Legacy and Influence

The discontinuation of the DEC Alpha architecture stemmed from corporate mergers and strategic shifts toward x86 compatibility. Following Digital Equipment Corporation's acquisition by in 1998, , a major customer, announced plans to phase out Alpha in favor of the and architecture. After acquired in 2001, the company accelerated the transition to x86-based systems, marking the end of new Alpha development. The final systems were sold until 2006, fulfilling contractual obligations while focused on and x86 platforms. Alpha's software ecosystem included notable ports and emulation efforts that extended its usability post-discontinuation. In the 1990s, Microsoft ported to Alpha, starting with version 3.1 in 1993—developed largely by DEC engineers—and continuing through NT 4.0 in 1996, with development ceasing in 1999 as focus shifted to x86. Modern preserves Alpha's legacy, with providing full system and user-mode support for the architecture since the early 2000s, enabling execution of Alpha binaries on contemporary hardware. In November 2025, the kernel's Alpha port gained a new maintainer, sustaining community-driven support for the architecture. Alpha's design principles influenced subsequent 64-bit RISC architectures, particularly through explicit lessons in explicit 64-bit addressing and superscalar pipelining. In 2001, Intel acquired Alpha's intellectual property from , incorporating elements like advanced branch prediction and techniques into Itanium's evolution, such as the Montecito and Tukwila processors. As one of the earliest commercial 64-bit RISC implementations, launched in 1992, Alpha contributed to broader trends in , informing the development of ARM64 and PowerPC64 by demonstrating scalable virtual addressing up to 43 bits and load/store simplicity without legacy CISC baggage. Patents related to Alpha's core innovations, filed in the early , began expiring in the , facilitating greater academic and hobbyist study of its . This, combined with the release of related and documentation into open-source repositories in the , has sustained interest. In the , Alpha maintains relevance in retro computing communities through for running historical operating systems like Tru64 UNIX and , alongside emerging FPGA-based recreations for educational hardware prototyping.

References

  1. [1]
    [PDF] Alpha Architecture - Handbook - Bitsavers.org
    Instruction Format Overview..... 1-4. Instruction Overview. 1-4. Instruction Set Characteristics. 1-6. Terminology and Conventions.
  2. [2]
    [PDF] Alpha AXP Architecture and Systems - Bitsavers.org
    Cover Design. The DECchip 21064, the first implementation of Digital's Alpha AXP computer architecture, is the world's fastest single-chip microprocessor.
  3. [3]
    [PDF] DECChip™ 21064–AA RISC Microprocessor Preliminary Data Sheet
    This document describes the 21064-AA RISC CPU microprocessor. The 21064-AA is the first of a family of microprocessors that implement the Digital Equipment ...
  4. [4]
    [PDF] the long Road to 64 Bits
    Jan 1, 2009 · DEC. DEC shipped 64-bit Alpha sys- tems in late 1992, with a 64-bit operating system, and by late 1994 was shipping servers with memories ...
  5. [5]
    [PDF] THE ALPHA 21264 MICROPROCESSOR
    Alpha microprocessors have been performance leaders since their introduction in 1992. The first generation 21064 and the later 211641,2 raised expectations ...
  6. [6]
    Understanding DEC Alpha: Architecture & Modern Solutions
    The DEC Alpha processor is the successor of the outdated VAX systems that were released in 1992. The VAX systems were 32-bit architecture, while the DEC ...
  7. [7]
    What Happened to Alpha? - ITPro Today
    In late August, Compaq announced that the company will no longer sell Alpha for the NT platform. Rumors circulated that Compaq had laid off 120 engineers.
  8. [8]
    [PDF] Managing technological leaps : a study of DEC's alpha design team
    Digital's management not only agreed to adopt the RISC architecture from MIPS, they also decided to cancel PRISM! ALPHA -. LIFE AFTER PRISM. The decision to ...<|control11|><|separator|>
  9. [9]
    [PDF] Guide to the Digital Equipment Corporation records, 1947-2002
    The Alpha architecture, introduced in 1992, was implemented as a line of microprocessors. DEC produced other technologies besides computers. It sold a wide ...
  10. [10]
    [PDF] A Historical Look at the VAX: The Economics of Microprocessors ...
    Jan 24, 2006 · The original study team was called the "RISCy VAX Task Force." The advanced development work was labeled "EVAX." When the program was ...
  11. [11]
    1992 | Timeline of Computer History
    DEC announces Alpha chip architecture ... Designed to replace the 32-bit VAX architecture, the Alpha is a 64-bit reduced instruction set computer (RISC) ...
  12. [12]
    [PDF] Alpha AXP Architecture and Systems - VMS Software
    The DECchip 21064, the first implementation of Digital's Alpha AXP computer architecture, is the world's fastest single-chip microprocessor. Represented on our ...
  13. [13]
    [PDF] Alpha Architecture Handbook
    The Alpha architecture facilitates pipelining multiple instances of the same operations because there are no special registers and no condition codes. The ...
  14. [14]
    [PDF] EV3 AND EV4 SPECIFICATION - DC227 and DC228 - Bitsavers.org
    The EV3 and EV4 chips are the first in a family of microprocessors that implement the ALPHA architecture. The information in this document is subject to ...
  15. [15]
    Alpha architecture and first implementation - IEEE Xplore
    EV4, the first implementation of the Alpha architecture, is a 200-MHz custom VLSI CPU with a peak issue rate of 400 MIPs. EV4 is implemented in Digital's ...Missing: integration | Show results with:integration
  16. [16]
    Alpha 21064 - Microarchitectures - DEC - WikiChip
    Aug 4, 2017 · Release Dates[edit]​​ Tape-out for the Alpha 21064 occurred on July 14, 1991. First parts were available on August 30 and a successful boot-up ...Missing: speed | Show results with:speed
  17. [17]
    DEC Digital Alpha 21164 Microprocessor - Peripheral
    The Alpha 21164 was introduced in January 1995 at 266 MHz. A 300 MHz version was introduced in March 1995. The final Alpha 21164, a 333 MHz version, was ...
  18. [18]
    Alpha AXP 21164 - ACM Digital Library
    the 21164 that we tested can be calculated as twice its. CPU clock rate in MHz. (300 MHz,. 600 MFLOPS). A version of the 21064 running at 150 MHz is used in ...
  19. [19]
    DETAILS START TO EMERGE ABOUT DEC's EV5 ALPHA ...
    Jun 22, 1994 · Digital Equipment Corp's second generation Alpha processor, the EV5, will have a higher performance on-chip cache than existing Alphas and a ...
  20. [20]
    [PDF] The Alpha 21264 Microprocessor Architecture | Manx Docs
    The Alpha 21264 is a super-scalar, third-generation microprocessor with out-of-order and speculative execution, a 600 MHz cycle time, and a high-performance ...
  21. [21]
    [PDF] Alpha 21364 to Ease Memory Bottleneck - CECS
    Oct 26, 1998 · With two levels of on-chip cache, the 21364 harkens back to the 21164 (see MPR 9/12/94, p. 1) in its cache design. Digital itself repudiated ...
  22. [22]
    Design tradeoffs for the alpha EV8 conditional branch predictor
    The Alpha EV8 microprocessor project, canceled in June 2001 in a late phase of development, envisioned an aggressive 8-wide issue out-of-order superscalar ...
  23. [23]
    Alpha AXP System Reference Manual
    This document describes the Alpha AXP architecture. This information shall not be disclosed to non-Digital personnel or generally distributed within Digital.
  24. [24]
    Compaq to buy Digital for $9.6 billion - CNET
    Jan 26, 1998 · Compaq likely will continue to sell Digital's Alpha and Intel-based workstations and only incrementally merge product lines. Like many other ...
  25. [25]
    Long gone, DEC is still powering the world of computing
    Oct 6, 2023 · In 1992, it introduced the Alpha AXP, later shortened to just Alpha, a RISC-based processor designed to compete with the other RISC chips on the ...
  26. [26]
    [PDF] Alpha Architecture Reference Manual
    Jun 1, 2010 · This document is a complete description of the Alpha architecture, derived from an internal manual, and covers the Alpha approach to RISC  ...Missing: predicate | Show results with:predicate
  27. [27]
    [PDF] Alpha Assembly Language Guide - Carnegie Mellon University
    Sep 22, 1998 · The Alpha architecture is a 64-bit RISC, with 64-bit registers. It uses 4-byte and 8-byte integers, and arithmetic operations are register- ...
  28. [28]
    [PDF] Digital Equipment Corporation ALPHA Calling Standard
    Apr 27, 1990 · This register contains a pointer to the top of the current operating stack. ... is the integer Read-As-Zero register, R30 is the hardware SP, and ...
  29. [29]
    The Alpha AXP, part 5: Conditional operations and control flow
    Aug 11, 2017 · These instructions store the address of the subsequent instruction (the return address) in the Ra register and then transfer to the destination.
  30. [30]
    [PDF] Alpha Architecture Handbook - O3ONE
    The Alpha architecture facilitates pipelining multiple instances of the same operations because there are no special registers and no condition codes. The ...
  31. [31]
    Motion Video Instructions - AlphaLinux
    Aug 29, 2019 · Beginning with the PCA56 processor, DEC added the Motion Video Instructions (MVI) to accelerate algorithms related to motion video formats ...
  32. [32]
    [PDF] DECchip 21164-AA (EV5 CPU) Functional Specification
    Oct 18, 1993 · On-chip write buffer with six 32-byte entries . On-chip 96Kbyte 3-way set associative writeback second level cache . Bus interface unit ...
  33. [33]
    Alpha AXP Legacy: Powerhouse for Data-Intensive Applications
    Byte Word Extension (BWX): These are the added instructions used to manipulate 8-bit and 16-bit quantities, elevating the performance of applications that don't ...
  34. [34]
    What is Alpha 21064 Processor? - GeeksforGeeks
    Mar 15, 2022 · Caches: The Alpha 21064 has two on-die primary caches named as I-Cache and D-Cache. The I-Cache is an 8 KB instruction cache while the D ...Missing: integration | Show results with:integration
  35. [35]
    [PDF] Alpha 21164 Microprocessor Data Sheet
    The 21164 microprocessor is a high-performance implementation of DIGITAL's. Alpha architecture. The following sections provide an overview of the chip's archi-.
  36. [36]
    [PDF] Superscalar instruction execution in the 21164 Alpha microprocessor
    1-mm' chip, which contains 9.3 million transistors. Key to the Alpha 21 164 architectural perfor- mance are its four-way superscalar instruction issue; low ...Missing: core | Show results with:core
  37. [37]
    [PDF] The Alpha 21264 Microprocessor: Out-of-Order Execution at 600 MHz
    Continued Alpha performance leadership. 600 MHz operation in 0.35u CMOS6, 6 metal layers, 2.2V. 15 Million transistors, 3.1 cm2, 587 pin PGA.Missing: EV6 1997
  38. [38]
    [PDF] Alpha 21264 Microprocessor Data Sheet - Index of /
    Maximum Ta for 21264 @ 600 MHz with Various Airflows. ... Alpha 21264 500-MHz microprocessor. 21264-A1. Title. Order Number. Alpha Architecture Reference ...Missing: 1997 | Show results with:1997
  39. [39]
    [PDF] HP AlphaServer/AlphaStation ES47 Tower ... - Island Computers
    At A Glance. AlphaServer ES47/ES80 systems. Up to 8 Alpha 21364 EV7 processors at 1150 MHz and 1000 MHz with advanced on-chip memory controllers and switch.
  40. [40]
    None
    ### Key Specifications for Alpha 21364 (EV7)
  41. [41]
    Design Tradeoffs for the Alpha EV8 Conditional Branch Predictor
    The Alpha EV8 microprocessor project, canceled in June 2001 in a late phase of development, envisioned an aggressive 8-wide issue out-of-order superscalar ...<|separator|>
  42. [42]
    Process Technology History - DEC - WikiChip
    Dec 26, 2017 · This article details details DEC's semiconductor process technology history for research and posterity. The table below shows the history of ...Missing: BiCMOS | Show results with:BiCMOS
  43. [43]
    Alpha 21364 - Wikipedia
    The Alpha 21364, code-named "Marvel", also known as EV7 is a microprocessor developed by Digital Equipment Corporation (DEC), later Compaq Computer Corporation.Missing: EV4 EV5 EV6
  44. [44]
    [PDF] Alpha 21364 (EV7) - WikiChip
    Jan 4, 2002 · 1.75 MB, 7-way set associative, with ECC ▪ 20 GB/s total read/write bandwidth ▪ 16 Victim buffers for L1 -> L2 ▪ 16 Victim buffers for L2 -> ...Missing: width | Show results with:width
  45. [45]
    [PDF] AlphaServer ES47, ES80, and GS1280 systems - UCSF RBVI
    The EV7 integrates on a single chip the logic and switching for building multi-processor systems of any size. The architecture enables all processors, memory, ...
  46. [46]
    [PDF] AlphaServer GS80, GS160, and GS320 Systems Technical Summary
    The two-level switch design provides for incremental growth, supported by a modular power system that is key to the system management control and redundant.Missing: EV7 | Show results with:EV7
  47. [47]
    Performance analysis of the Alpha 21364-based HP GS1280 ...
    This paper evaluates performance characteristics of the HP GS1280 shared memory multiprocessor system. The GS1280 system contains up to 64 Alpha 21364 CPUs ...
  48. [48]
    Samsung Licensed to Market Digital's Alpha Microprocessors
    Jun 19, 1996 · The Digital Equipment Corporation said today that it had licensed Samsung Electronics to make and market its Alpha microprocessors worldwide ...
  49. [49]
    Samsung Commences Alpha 21164 Chip Test Production - HPCwire
    Nov 15, 1996 · Samsung said it will begin mass production of the Alpha 21164 family in the second quarter of 1997 at its advanced eight-inch fabrication plant ...
  50. [50]
    Samsung starts up 600-MHz Alpha - CNET
    Nov 30, 1998 · Samsung Electronics in Korea is ready to mass produce the third generation of Alpha processors, which compete with Intel chips, ...<|separator|>
  51. [51]
    Samsung has rights to future, unique Alpha chips | ZDNET
    Feb 10, 1998 · Digital said last night it had granted Samsung a licence allowing the Korean firm to make its own versions of the Alpha processor.
  52. [52]
    DEC Alpha | Microsoft Wiki | Fandom
    The first version, the Alpha 21064 or EV4, was the first CMOS microprocessor whose operating frequency rivalled higher-powered ECL minicomputers and mainframes.Missing: timeline | Show results with:timeline
  53. [53]
    AlphaStation 200 4/166 - Computer History Wiki
    Oct 29, 2024 · AlphaStation 200 4/166. Summary. Announcement date: 1 November 1994. OS support (VMS):, OpenVMS V6.1-1H1. CPU Details.
  54. [54]
  55. [55]
    Evaluation of a Commercial Microprocessor - ACM Digital Library
    Two SPARC microprocessors were started in the early 1990s: the Fujitsu-Ross ... four microprocessor families, the DEC Alpha, the Intel Pentium, the HP PA-RISC, ...
  56. [56]
    Digital DEC Multia (Alpha Generation) - Computer
    The Multia, later rebranded the Universal Desktop Box, was a line of desktop computers introduced by Digital Equipment Corporation on 7 November 1994.Missing: one | Show results with:one
  57. [57]
    [PDF] Digital AlphaStation 200/400 Series Technical Information
    Internal Bay Availability: AlphaStation 200 Series Systems ... address in the AlphaStation 200 Series and the AlphaStation 400 Series systems.Missing: 1994 | Show results with:1994
  58. [58]
    [PDF] AlphaServer 1000A
    The Digital AlphaServer 1000A system is a low-cost, single- processor, PCI/EISA-based server. It is suitable for general-.
  59. [59]
    [PDF] AlphaServer 4000/4100 Service Manual - Manx Docs
    The AlphaServer 4100 system bus connects up to four CPUs, four pairs of memory ... power-up test flow, 2-8 tests, 2-10. Standard I/O, 1-32 status command ...
  60. [60]
    [PDF] AlphaServer DS20 User's Guide - Manx Docs
    This manual is for anyone who manages, operates, or services the. Compaq AlphaServer DS20 system. It covers operation, firmware, initial troubleshooting, and ...
  61. [61]
    Tru64 UNIX - EmuVM
    Oct 25, 2021 · Support continued until 2012. Our product AlphaVM-Pro is an Alpha hardware emulator that allows to create a virtual AlphaServer system. It is ...
  62. [62]
    HP OpenVMS Cluster Software
    Some of the new generation AlphaServer processors will support DSSI. The GS series and the DS20 series will have support. Other DS series and the ES series will ...
  63. [63]
    DEC Alpha - Computer History Wiki
    Nov 15, 2024 · Alpha, originally known as Alpha AXP, is a 64-bit Reduced Instruction Set Computer (RISC) instruction set architecture (ISA) developed by Digital Equipment ...Missing: models | Show results with:models
  64. [64]
    [PDF] Digital Plans Broad Alpha Processor Family - CECS
    Nov 18, 1992 · SPECint92 and 126.0 SPECfp92, while the 200-MHz. DEC 10000 reaches 106.5 SPECint92 and 200.4. SPECfp92. The Model 400, a desktop system, is ...
  65. [65]
    [PDF] Alpha XL 300/366 Workstations
    300MHz EV5 32-512MB. 17" C. SIMM. 32MB 2MB RX,CD 1GB NI. Matrox Millenium 2D ... SPECint95 SPECfp95. 7.3. 9.8. 12.2. 13.4. 300MHz. 366MHz. RZxx. RZ26L-VW 1GB.
  66. [66]
    CFP2000 Result: Alpha Processor, Inc. API UP2000 833 MHz
    Benchmark, Reference Time, Base Runtime, Base Ratio, Runtime, Ratio, Graph Scale. 168.wupwise, 1600, 287, 557, 261, 614, 168.wupwise base result bar (557)
  67. [67]
    [PS] Performance of Various Computers Using Standard Linear ...
    “LINPACK Benchmark”. OS/Compiler. n=100. Mflop/s. “TPP”. Best. Effort. n=1000. Mflop/s. “Theoritical. Peak”. Mflop/s. Compaq/DEC Alpha 21264 EV67 500 MHz. -O5 - ...
  68. [68]
    Modern Microprocessors - A 90-Minute Guide! - Lighterra
    A 200 MHz MIPS R10000, a 300 MHz UltraSPARC and a 400 MHz Alpha 21164 were all about the same speed at running most programs, yet they differed by a factor of ...
  69. [69]
    How much better was DEC Alpha than contemporary x86? - Quora
    Feb 7, 2020 · Is the DEC Alpha CPU in 1996 faster than today's Intel CPU? The highest core frequency of DEC Alphas back in ...Missing: condition | Show results with:condition
  70. [70]
    DEC Alpha: Understanding Its Core and Differences with x86
    After acquiring DEC, Compaq decided to phase out the Alpha architecture and ultimately selected the Itanium architecture developed in partnership with Hewlett- ...Missing: 2004 | Show results with:2004
  71. [71]
    HP Rolls Out 64-Bit AlphaServer - Channel Insider
    Oct 20, 2003 · HP has rolled out a road map in which the company will introduce the last new Alpha chip in 2004 and will stop selling AlphaServers in 2006.
  72. [72]
    The Death of Alpha on NT - ITPro Today
    The Alpha on NT story has its roots back to the inception of NT. Dave Cutler, NT's creator, was working on a new OS, code-named "Mica," for Digital Equipment.
  73. [73]
    Emulation — QEMU documentation
    Both System Emulation and User Mode Emulation are supported depending on the guest architecture. Supported Guest Architectures for Emulation . Architecture ...
  74. [74]
    Itanium to take on Alpha influence - CNET
    Jan 30, 2002 · The Alpha influence in these chips comes as a result of a deal last June between Intel and Compaq under which Intel acquired a license for the ...
  75. [75]
    [PDF] Alpha AXP Architecture - Dan Luu
    Richard L. Sites Dick Sites is a senior consul- tant engineer in the Semiconductor Engineering. Group, where he is working on binary translators and the ...