ARM DMIPS算力说明

ARM DMIPS算力说明

ARM算力参考官网地址

https://en.wikipedia.org/wiki/List_of_ARM_processors

Product familyARM architectureProcessorFeatureCache (I / D), MMUTypical MIPS @ MHzReference
ARM1ARMv1ARM1First implementationNone
ARM2ARMv2ARM2ARMv2 added the MUL (multiply) instructionNone0.33 DMIPS/MHz
ARM2aSARMv2aARM250Integrated MEMC (MMU), graphics and I/O processor. ARMv2a added the SWP and SWPB (swap) instructionsNone, MEMC1a
ARM3First integrated memory cache4 KB unified0.50 DMIPS/MHz
ARM6ARMv3ARM60ARMv3 first to support 32-bit memory address space (previously 26-bit).
ARMv3M first added long multiply instructions (32x32=64).
None10 MIPS @ 12 MHz
ARM600As ARM60, cache and coprocessor bus (for FPA10 floating-point unit)4 KB unified28 MIPS @ 33 MHz
ARM610As ARM60, cache, no coprocessor bus4 KB unified17 MIPS @ 20 MHz
0.65 DMIPS/MHz
[4]
ARM7ARMv3ARM700coprocessor bus (for FPA11 floating-point unit)8 KB unified40 MHz
ARM710As ARM700, no coprocessor bus8 KB unified40 MHz[5]
ARM710aAs ARM710, also used as core of ARM71008 KB unified40 MHz
0.68 DMIPS/MHz
ARM7TARMv4TARM7TDMI(-S)3-stage pipeline, Thumb, ARMv4 first to drop legacy ARM 26-bit addressingNone15 MIPS @ 16.8 MHz
63 DMIPS @ 70 MHz
ARM710TAs ARM7TDMI, cache8 KB unified, MMU36 MIPS @ 40 MHz
ARM720TAs ARM7TDMI, cache8 KB unified, MMU with FCSE (Fast Context Switch Extension)60 MIPS @ 59.8 MHz
ARM740TAs ARM7TDMI, cacheMPU
ARM7EJARMv5TEJARM7EJ-S5-stage pipeline, Thumb, Jazelle DBX, enhanced DSP instructionsNone
ARM8ARMv4ARM8105-stage pipeline, static branch prediction, double-bandwidth memory8 KB unified, MMU84 MIPS @ 72 MHz
1.16 DMIPS/MHz
[6][7]
ARM9TARMv4TARM9TDMI5-stage pipeline, ThumbNone
ARM920TAs ARM9TDMI, cache16 KB / 16 KB, MMU with FCSE (Fast Context Switch Extension)200 MIPS @ 180 MHz[8]
ARM922TAs ARM9TDMI, caches8 KB / 8 KB, MMU
ARM940TAs ARM9TDMI, caches4 KB / 4 KB, MPU
ARM9EARMv5TEARM946E-SThumb, enhanced DSP instructions, cachesVariable, tightly coupled memories, MPU
ARM966E-SThumb, enhanced DSP instructionsNo cache, TCMs
ARM968E-SAs ARM966E-SNo cache, TCMs
ARMv5TEJARM926EJ-SThumb, Jazelle DBX, enhanced DSP instructionsVariable, TCMs, MMU220 MIPS @ 200 MHz
ARMv5TEARM996HSClockless processor, as ARM966E-SNo caches, TCMs, MPU
ARM10EARMv5TEARM1020E6-stage pipeline, Thumb, enhanced DSP instructions, (VFP)32 KB / 32 KB, MMU
ARM1022EAs ARM1020E16 KB / 16 KB, MMU
ARMv5TEJARM1026EJ-SThumb, Jazelle DBX, enhanced DSP instructions, (VFP)Variable, MMU or MPU
ARM11ARMv6ARM1136J(F)-S8-stage pipeline, SIMD, Thumb, Jazelle DBX, (VFP), enhanced DSP instructions, unaligned memory accessVariable, MMU740 @ 532–665 MHz (i.MX31 SoC), 400–528 MHz[9]
ARMv6T2ARM1156T2(F)-S9-stage pipeline, SIMD, Thumb-2, (VFP), enhanced DSP instructionsVariable, MPU[10]
ARMv6ZARM1176JZ(F)-SAs ARM1136EJ(F)-SVariable, MMU + TrustZone965 DMIPS @ 772 MHz, up to 2,600 DMIPS with four processors[11]
ARMv6KARM11MPCoreAs ARM1136EJ(F)-S, 1–4 core SMPVariable, MMU
SecurCoreARMv6-MSC000As Cortex-M00.9 DMIPS/MHz
ARMv4TSC100As ARM7TDMI
ARMv7-MSC300As Cortex-M31.25 DMIPS/MHz
Cortex-MARMv6-MCortex-M0Microcontroller profile, most Thumb + some Thumb-2,[12] hardware multiply instruction (optional small), optional system timer, optional bit-banding memoryOptional cache, no TCM, no MPU0.84 DMIPS/MHz[13]
Cortex-M0+Microcontroller profile, most Thumb + some Thumb-2,[12] hardware multiply instruction (optional small), optional system timer, optional bit-banding memoryOptional cache, no TCM, optional MPU with 8 regions0.93 DMIPS/MHz[14]
Cortex-M1Microcontroller profile, most Thumb + some Thumb-2,[12] hardware multiply instruction (optional small), OS option adds SVC / banked stack pointer, optional system timer, no bit-banding memoryOptional cache, 0–1024 KB I-TCM, 0–1024 KB D-TCM, no MPU136 DMIPS @ 170 MHz,[15] (0.8 DMIPS/MHz FPGA-dependent)[16][17]
ARMv7-MCortex-M3Microcontroller profile, Thumb / Thumb-2, hardware multiply and divide instructions, optional bit-banding memoryOptional cache, no TCM, optional MPU with 8 regions1.25 DMIPS/MHz[18]
ARMv7E-MCortex-M4Microcontroller profile, Thumb / Thumb-2 / DSP / optional VFPv4-SP single-precision FPU, hardware multiply and divide instructions, optional bit-banding memoryOptional cache, no TCM, optional MPU with 8 regions1.25 DMIPS/MHz (1.27 w/FPU)[19]
Cortex-M7Microcontroller profile, Thumb / Thumb-2 / DSP / optional VFPv5 single and double precision FPU, hardware multiply and divide instructions0−64 KB I-cache, 0−64 KB D-cache, 0–16 MB I-TCM, 0–16 MB D-TCM (all these w/optional ECC), optional MPU with 8 or 16 regions2.14 DMIPS/MHz[20]
ARMv8-M BaselineCortex-M23Microcontroller profile, Thumb-1 (most), Thumb-2 (some), Divide, TrustZoneOptional cache, no TCM, optional MPU with 16 regions1.03 DMIPS/MHz[21]
ARMv8-M MainlineCortex-M33Microcontroller profile, Thumb-1, Thumb-2, Saturated, DSP, Divide, FPU (SP), TrustZone, Co-processorOptional cache, no TCM, optional MPU with 16 regions1.50 DMIPS/MHz[22]
Cortex-M35PMicrocontroller profile, Thumb-1, Thumb-2, Saturated, DSP, Divide, FPU (SP), TrustZone, Co-processorBuilt-in cache (with option 2–16 KB), I-cache, no TCM, optional MPU with 16 regions1.50 DMIPS/MHz[23]
ARMv8.1-M MainlineCortex-M521.60 DMIPS/MHz[24]
Cortex-M551.69 DMIPS/MHz[25]
Cortex-M853.13 DMIPS/MHz[26]
Cortex-RARMv7-RCortex-R4Real-time profile, Thumb / Thumb-2 / DSP / optional VFPv3 FPU, hardware multiply and optional divide instructions, optional parity & ECC for internal buses / cache / TCM, 8-stage pipeline dual-core running lockstep with fault logic0–64 KB / 0–64 KB, 0–2 of 0–8 MB TCM, opt. MPU with 8/12 regions1.67 DMIPS/MHz[27][28]
Cortex-R5Real-time profile, Thumb / Thumb-2 / DSP / optional VFPv3 FPU and precision, hardware multiply and optional divide instructions, optional parity & ECC for internal buses / cache / TCM, 8-stage pipeline dual-core running lock-step with fault logic / optional as 2 independent cores, low-latency peripheral port (LLPP), accelerator coherency port (ACP)[29]0–64 KB / 0–64 KB, 0–2 of 0–8 MB TCM, opt. MPU with 12/16 regions1.67 DMIPS/MHz[27][30]
Cortex-R7Real-time profile, Thumb / Thumb-2 / DSP / optional VFPv3 FPU and precision, hardware multiply and optional divide instructions, optional parity & ECC for internal buses / cache / TCM, 11-stage pipeline dual-core running lock-step with fault logic / out-of-order execution / dynamic register renaming / optional as 2 independent cores, low-latency peripheral port (LLPP), ACP[29]0–64 KB / 0–64 KB, ? of 0–128 KB TCM, opt. MPU with 16 regions2.50 DMIPS/MHz[27][31]
Cortex-R8TBD0–64 KB / 0–64 KB L1, 0–1 / 0–1 MB TCM, opt MPU with 24 regions2.50 DMIPS/MHz[27][32]
ARMv8-RCortex-R52TBD0–32 KB / 0–32 KB L1, 0–1 / 0–1 MB TCM, opt MPU with 24+24 regions2.09 DMIPS/MHz[33]
Cortex-R52+TBD0–32 KB / 0–32 KB L1, 0–1 / 0–1 MB TCM, opt MPU with 24+24 regions2.09 DMIPS/MHz[34]
Cortex-R82TBD16–128 KB /16–64 KB L1, 64K–1MB L2, 0.16–1 / 0.16–1 MB TCM,

opt MPU with 32+32 regions

3.41 DMIPS/MHz[35][36]
Cortex-A(32-bit)ARMv7-ACortex-A5Application profile, ARM / Thumb / Thumb-2 / DSP / SIMD / Optional VFPv4-D16 FPU / Optional NEON / Jazelle RCT and DBX, 1–4 cores / optional MPCore, snoop control unit (SCU), generic interrupt controller (GIC), accelerator coherence port (ACP)4−64 KB / 4−64 KB L1, MMU + TrustZone1.57 DMIPS/MHz per core[37]
Cortex-A7Application profile, ARM / Thumb / Thumb-2 / DSP / VFPv4 FPU / NEON / Jazelle RCT and DBX / Hardware virtualization, in-order execution, superscalar, 1–4 SMP cores, MPCore, Large Physical Address Extensions (LPAE), snoop control unit (SCU), generic interrupt controller (GIC), architecture and feature set are identical to A15, 8–10 stage pipeline, low-power design[38]8−64 KB / 8−64 KB L1, 0–1 MB L2, MMU + TrustZone1.9 DMIPS/MHz per core[39]
Cortex-A8Application profile, ARM / Thumb / Thumb-2 / VFPv3 FPU / NEON / Jazelle RCT and DAC, 13-stage superscalar pipeline16–32 KB / 16–32 KB L1, 0–1 MB L2 opt. ECC, MMU + TrustZoneUp to 2000 (2.0 DMIPS/MHz in speed from 600 MHz to greater than 1 GHz)[40]
Cortex-A9Application profile, ARM / Thumb / Thumb-2 / DSP / Optional VFPv3 FPU / Optional NEON / Jazelle RCT and DBX, out-of-order speculative issue superscalar, 1–4 SMP cores, MPCore, snoop control unit (SCU), generic interrupt controller (GIC), accelerator coherence port (ACP)16–64 KB / 16–64 KB L1, 0–8 MB L2 opt. parity, MMU + TrustZone2.5 DMIPS/MHz per core, 10,000 DMIPS @ 2 GHz on Performance Optimized TSMC 40G (dual-core)[41]
Cortex-A12Application profile, ARM / Thumb-2 / DSP / VFPv4 FPU / NEON / Hardware virtualization, out-of-order speculative issue superscalar, 1–4 SMP cores, Large Physical Address Extensions (LPAE), snoop control unit (SCU), generic interrupt controller (GIC), accelerator coherence port (ACP)32−64 KB3.0 DMIPS/MHz per core[42]
Cortex-A15Application profile, ARM / Thumb / Thumb-2 / DSP / VFPv4 FPU / NEON / integer divide / fused MAC / Jazelle RCT / hardware virtualization, out-of-order speculative issue superscalar, 1–4 SMP cores, MPCore, Large Physical Address Extensions (LPAE), snoop control unit (SCU), generic interrupt controller (GIC), ACP, 15-24 stage pipeline[38]32 KB w/parity / 32 KB w/ECC L1, 0–4 MB L2, L2 has ECC, MMU + TrustZoneAt least 3.5 DMIPS/MHz per core (up to 4.01 DMIPS/MHz depending on implementation)[43][44]
Cortex-A17Application profile, ARM / Thumb / Thumb-2 / DSP / VFPv4 FPU / NEON / integer divide / fused MAC / Jazelle RCT / hardware virtualization, out-of-order speculative issue superscalar, 1–4 SMP cores, MPCore, Large Physical Address Extensions (LPAE), snoop control unit (SCU), generic interrupt controller (GIC), ACP32 KB L1, 256 KB–8 MB L2 w/optional ECC2.8 DMIPS/MHz[45]
ARMv8-ACortex-A32Application profile, AArch32, 1–4 SMP cores, TrustZone, NEON advanced SIMD, VFPv4, hardware virtualization, dual issue, in-order pipeline8–64 KB w/optional parity / 8−64 KB w/optional ECC L1 per core, 128 KB–1 MB L2 w/optional ECC shared[46]
Cortex-A(64-bit)ARMv8-ACortex-A34Application profile, AArch64, 1–4 SMP cores, TrustZone, NEON advanced SIMD, VFPv4, hardware virtualization, 2-width decode, in-order pipeline8−64 KB w/parity / 8−64 KB w/ECC L1 per core, 128 KB–1 MB L2 shared, 40-bit physical addresses[47]
Cortex-A35Application profile, AArch32 and AArch64, 1–4 SMP cores, TrustZone, NEON advanced SIMD, VFPv4, hardware virtualization, 2-width decode, in-order pipeline8−64 KB w/parity / 8−64 KB w/ECC L1 per core, 128 KB–1 MB L2 shared, 40-bit physical addresses1.78 DMIPS/MHz[48]
Cortex-A53Application profile, AArch32 and AArch64, 1–4 SMP cores, TrustZone, NEON advanced SIMD, VFPv4, hardware virtualization, 2-width decode, in-order pipeline8−64 KB w/parity / 8−64 KB w/ECC L1 per core, 128 KB–2 MB L2 shared, 40-bit physical addresses2.3 DMIPS/MHz[49]
Cortex-A57Application profile, AArch32 and AArch64, 1–4 SMP cores, TrustZone, NEON advanced SIMD, VFPv4, hardware virtualization, 3-width decode superscalar, deeply out-of-order pipeline48 KB w/DED parity / 32 KB w/ECC L1 per core; 512 KB–2 MB L2 shared w/ECC; 44-bit physical addresses4.1–4.8 DMIPS/MHz[50][51][52]
Cortex-A72Application profile, AArch32 and AArch64, 1–4 SMP cores, TrustZone, NEON advanced SIMD, VFPv4, hardware virtualization, 3-width superscalar, deeply out-of-order pipeline48 KB w/DED parity / 32 KB w/ECC L1 per core; 512 KB–2 MB L2 shared w/ECC; 44-bit physical addresses6.3-7.3 DMIPS/MHz[53][54]
Cortex-A73Application profile, AArch32 and AArch64, 1–4 SMP cores, TrustZone, NEON advanced SIMD, VFPv4, hardware virtualization, 2-width superscalar, deeply out-of-order pipeline64 KB / 32−64 KB L1 per core, 256 KB–8 MB L2 shared w/ optional ECC, 44-bit physical addresses7.4-8.5 DMIPS/MHz[53][55]
ARMv8.2-ACortex-A55Application profile, AArch32 and AArch64, 1–8 SMP cores, TrustZone, NEON advanced SIMD, VFPv4, hardware virtualization, 2-width decode, in-order pipeline[56]16−64 KB / 16−64 KB L1, 256 KB L2 per core, 4 MB L3 shared3 DMIPS/MHz[53][57]
Cortex-A65Application profile, AArch64, 1–8 SMP cores, TrustZone, NEON advanced SIMD, VFPv4, hardware virtualization, 2-wide decode superscalar, 3-width issue, out-of-order pipeline, SMT[58]
Cortex-A65AEAs ARM Cortex-A65, adds dual core lockstep for safety applications64 / 64 KB L1, 256 KB L2 per core, 4 MB L3 shared[59]
Cortex-A75Application profile, AArch32 and AArch64, 1–8 SMP cores, TrustZone, NEON advanced SIMD, VFPv4, hardware virtualization, 3-width decode superscalar, deeply out-of-order pipeline[60]64 / 64 KB L1, 512 KB L2 per core, 4 MB L3 shared8.2-9.5 DMIPS/MHz[53][61]
Cortex-A76Application profile, AArch32 (non-privileged level or EL0 only) and AArch64, 1–4 SMP cores, TrustZone, NEON advanced SIMD, VFPv4, hardware virtualization, 4-width decode superscalar, 8-way issue, 13 stage pipeline, deeply out-of-order pipeline[62]64 / 64 KB L1, 256−512 KB L2 per core, 512 KB−4 MB L3 shared10.7-12.4 DMIPS/MHz[53][63]
Cortex-A76AEAs ARM Cortex-A76, adds dual core lockstep for safety applications[64]
Cortex-A77Application profile, AArch32 (non-privileged level or EL0 only) and AArch64, 1–4 SMP cores, TrustZone, NEON advanced SIMD, VFPv4, hardware virtualization, 4-width decode superscalar, 6-width instruction fetch, 12-way issue, 13 stage pipeline, deeply out-of-order pipeline[62]1.5K L0 MOPs cache, 64 / 64 KB L1, 256−512 KB L2 per core, 512 KB−4 MB L3 shared13-16 DMIPS/MHz[65][66]
Cortex-A78[67]
Cortex-A78AEAs ARM Cortex-A78, adds dual core lockstep for safety applications[68]
Cortex-A78C[69]
ARMv9-ACortex-A510
Cortex-A710[70]
Cortex-A715
Cortex-A520
Cortex-A720
Cortex-XARMv8.2-ACortex-X1Performance-tuned variant of Cortex-A78
ARMv9-ACortex-X2
Cortex-X3
Cortex-X4
NeoverseARMv8.2-ANeoverse N1Application profile, AArch32 (non-privileged level or EL0 only) and AArch64, 1–4 SMP cores, TrustZone, NEON advanced SIMD, VFPv4, hardware virtualization, 4-width decode superscalar, 8-way dispatch/issue, 13 stage pipeline, deeply out-of-order pipeline[62]64 / 64 KB L1, 512−1024 KB L2 per core, 2−128 MB L3 shared, 128 MB system level cache[71]
Neoverse E1Application profile, AArch64, 1–8 SMP cores, TrustZone, NEON advanced SIMD, VFPv4, hardware virtualization, 2-wide decode superscalar, 3-width issue, 10 stage pipeline, out-of-order pipeline, SMT32−64 KB / 32−64 KB L1, 256 KB L2 per core, 4 MB L3 shared[72]
ARMv8.4-ANeoverse V1[73]
ARMv9-ANeoverse N2[74]
ARM familyARM architectureARM coreFeatureCache (I / D), MMUTypical MIPS @ MHzReference

Designed by third parties[edit]

These cores implement the ARM instruction set, and were developed independently by companies with an architectural license from ARM.

Product familyARM architectureProcessorFeatureCache (I / D), MMUTypical MIPS @ MHz
StrongARM
(Digital)
ARMv4SA-1105-stage pipeline16 KB / 16 KB, MMU100–233 MHz
1.0 DMIPS/MHz
SA-1100derivative of the SA-11016 KB / 8 KB, MMU
Faraday[75]
(Faraday Technology)
ARMv4FA5106-stage pipelineUp to 32 KB / 32 KB cache, MPU1.26 DMIPS/MHz
100–200 MHz
FA526Up to 32 KB / 32 KB cache, MMU1.26 MIPS/MHz
166–300 MHz
FA6268-stage pipeline32 KB / 32 KB cache, MMU1.35 DMIPS/MHz
500 MHz
ARMv5TEFA606TE5-stage pipelineNo cache, no MMU1.22 DMIPS/MHz
200 MHz
FA626TE8-stage pipeline32 KB / 32 KB cache, MMU1.43 MIPS/MHz
800 MHz
FMP626TE8-stage pipeline, SMP1.43 MIPS/MHz
500 MHz
FA726TE13 stage pipeline, dual issue2.4 DMIPS/MHz
1000 MHz
XScale
(Intel / Marvell)
ARMv5TEXScale7-stage pipeline, Thumb, enhanced DSP instructions32 KB / 32 KB, MMU133–400 MHz
BulverdeWireless MMX, wireless SpeedStep added32 KB / 32 KB, MMU312–624 MHz
Monahans[76]Wireless MMX2 added32 KB / 32 KB L1, optional L2 cache up to 512 KB, MMUUp to 1.25 GHz
Sheeva
(Marvell)
ARMv5Feroceon5–8 stage pipeline, single-issue16 KB / 16 KB, MMU600–2000 MHz
Jolteon5–8 stage pipeline, dual-issue32 KB / 32 KB, MMU
PJ1 (Mohawk)5–8 stage pipeline, single-issue, Wireless MMX232 KB / 32 KB, MMU1.46 DMIPS/MHz
1.06 GHz
ARMv6 / ARMv7-APJ46–9 stage pipeline, dual-issue, Wireless MMX2, SMP32 KB / 32 KB, MMU2.41 DMIPS/MHz
1.6 GHz
Snapdragon
(Qualcomm)
ARMv7-AScorpion[77]1 or 2 cores. ARM / Thumb / Thumb-2 / DSP / SIMD / VFPv3 FPU / NEON (128-bit wide)256 KB L2 per core2.1 DMIPS/MHz per core
Krait[77]1, 2, or 4 cores. ARM / Thumb / Thumb-2 / DSP / SIMD / VFPv4 FPU / NEON (128-bit wide)4 KB / 4 KB L0, 16 KB / 16 KB L1, 512 KB L2 per core3.3 DMIPS/MHz per core
ARMv8-AKryo[78]4 cores.?Up to 2.2 GHz

(6.3 DMIPS/MHz)

Ax
(Apple)
ARMv7-ASwift[79]2 cores. ARM / Thumb / Thumb-2 / DSP / SIMD / VFPv4 FPU / NEONL1: 32 KB / 32 KB, L2: 1 MB shared3.5 DMIPS/MHz per core
ARMv8-ACyclone[80]2 cores. ARM / Thumb / Thumb-2 / DSP / SIMD / VFPv4 FPU / NEON / TrustZone / AArch64. Out-of-order, superscalar.L1: 64 KB / 64 KB, L2: 1 MB shared
SLC: 4 MB
1.3 or 1.4 GHz
ARMv8-ATyphoon[80][81]2 or 3 cores. ARM / Thumb / Thumb-2 / DSP / SIMD / VFPv4 FPU / NEON / TrustZone / AArch64L1: 64 KB / 64 KB, L2: 1 MB or 2 MB shared
SLC: 4 MB
1.4 or 1.5 GHz
ARMv8-ATwister[82]2 cores. ARM / Thumb / Thumb-2 / DSP / SIMD / VFPv4 FPU / NEON / TrustZone / AArch64L1: 64 KB / 64 KB, L2: 2 MB shared
SLC: 4 MB or 0 MB
1.85 or 2.26 GHz
ARMv8-AHurricane and Zephyr[83]Hurricane: 2 or 3 cores. AArch64, out-of-order, superscalar, 6-decode, 6-issue, 9-wide
Zephyr: 2 or 3 cores. AArch64, out-of-order, superscalar.
L1: 64 KB / 64 KB, L2: 3 MB or 8 MB shared
L1: 32 KB / 32 KB. L2: none
SLC: 4 MB or 0 MB
2.34 or 2.38 GHz
1.05 GHz
ARMv8.2-AMonsoon and Mistral[84]Monsoon: 2 cores. AArch64, out-of-order, superscalar, 7-decode, ?-issue, 11-wide
Mistral: 4 cores. AArch64, out-of-order, superscalar. Based on Swift.
L1I: 128 KB, L1D: 64 KB, L2: 8 MB shared
L1: 32 KB / 32 KB, L2: 1 MB shared
SLC: 4 MB
2.39 GHz
1.70 GHz
ARMv8.3-AVortex and Tempest[85]Vortex: 2 or 4 cores. AArch64, out-of-order, superscalar, 7-decode, ?-issue, 11-wide
Tempest: 4 cores. AArch64, out-of-order, superscalar, 3-decode. Based on Swift.
L1: 128 KB / 128 KB, L2: 8 MB shared
L1: 32 KB / 32 KB, L2: 2 MB shared
SLC: 8 MB
2.49 GHz
1.59 GHz
ARMv8.4-ALightning and Thunder[86]Lightning: 2 cores. AArch64, out-of-order, superscalar, 7-decode, ?-issue, 11-wide
Thunder: 4 cores. AArch64, out-of-order, superscalar.
L1: 128 KB / 128 KB, L2: 8 MB shared
L1: 32 KB / 48 KB, L2: 4 MB shared
SLC: 16 MB
2.66 GHz
1.73 GHz
ARMv8.5-AFirestorm and Icestorm[87]Firestorm: 2 cores. AArch64, out-of-order, superscalar, 8-decode, ?-issue, 14-wide
Icestorm: 4 cores. AArch64, out-of-order, superscalar, 4-decode, ?-issue, 7-wide.
L1: 192 KB / 128 KB, L2: 8 MB shared
L1: 128 KB / 64 KB, L2: 4 MB shared
SLC: 16 MB
3.0 GHz
1.82 GHz
ARMv8.5-AAvalanche and BlizzardAvalanche: 2 cores. AArch64, out-of-order, superscalar, 8-decode, ?-issue, 14-wide
Blizzard: 4 cores. AArch64, out-of-order, superscalar, 4-decode, ?-issue, 8-wide.
L1: 192 KB / 128 KB, L2: 12 MB shared
L1: 128 KB / 64 KB, L2: 4 MB shared
SLC: 32 MB
2.93 or 3.23 GHz
2.02 GHz
ARMv8.5-AEverest and SawtoothEverest: 2 cores. AArch64, out-of-order, superscalar, 8-decode, ?-issue, 14-wide
Sawtooth: 4 cores. AArch64, out-of-order, superscalar, 4-decode, ?-issue, 8-wide.
L1: 192 KB / 128 KB, L2: 16 MB shared
L1: 128 KB / 64 KB, L2: 4 MB shared
SLC: 24 MB
3.46 GHz
2.02 GHz
Mx
(Apple)
ARMv8.4-AFirestorm and IcestormFirestorm: 4, 6, 8 or 16 cores. AArch64, out-of-order, superscalar, 8-decode, ?-issue, 14-wide
Icestorm: 2 or 4 cores. AArch64, out-of-order, superscalar, 4-decode, ?-issue, 7-wide.
L1: 192 KB / 128 KB, L2: 12, 24 or 48 MB shared
L1: 128 KB / 64 KB, L2: 4 or 8 MB shared
SLC: 8, 24, 48 or 96 MB
3.2-3.23 GHz
2.06 GHz
ARMv8.5-AAvalanche and BlizzardAvalanche: 4, 6, 8 or 16 cores. AArch64, out-of-order, superscalar, 8-decode, ?-issue, 14-wide
Blizzard: 4 or 8 cores. AArch64, out-of-order, superscalar, 4-decode, ?-issue, 8-wide.
L1: 192 KB / 128 KB, L2: 16, 32 or 64 MB shared
L1: 128 KB / 64 KB, L2: 4 or 8 MB shared
SLC: 8, 24, 48 or 96 MB
3.49 GHz
2.42 GHz
X-Gene
(Applied Micro)
ARMv8-AX-Gene64-bit, quad issue, SMP, 64 cores[88]Cache, MMU, virtualization3 GHz (4.2 DMIPS/MHz per core)
Denver
(Nvidia)
ARMv8-ADenver[89][90]2 cores. AArch64, 7-wide superscalar, in-order, dynamic code optimization, 128 MB optimization cache,
Denver1: 28 nm, Denver2:16 nm
128 KB I-cache / 64 KB D-cacheUp to 2.5 GHz
Carmel
(Nvidia)
ARMv8.2-ACarmel[91][92]2 cores. AArch64, 10-wide superscalar, in-order, dynamic code optimization, ? MB optimization cache,
functional safety, dual execution, parity & ECC
? KB I-cache / ? KB D-cacheUp to ? GHz
ThunderX
(Cavium)
ARMv8-AThunderX64-bit, with two models with 8–16 or 24–48 cores (×2 w/two chips)?Up to 2.2 GHz
K12
(AMD)
ARMv8-AK12[93]???
Exynos
(Samsung)
ARMv8-AM1 ("Mongoose")[94]4 cores. AArch64, 4-wide, quad-issue, superscalar, out-of-order64 KB I-cache / 32 KB D-cache, L2: 16-way shared 2 MB5.1 DMIPS/MHz

(2.6 GHz)

ARMv8-AM2 ("Mongoose")4 cores. AArch64, 4-wide, quad-issue, superscalar, out-of-order64 KB I-cache / 32 KB D-cache, L2: 16-way shared 2 MB2.3 GHz
ARMv8-AM3 ("Meerkat")[95]4 cores, AArch64, 6-decode, 6-issue, 6-wide. superscalar, out-of-order64 KB I-cache / 64 KB D-cache, L2: 8-way private 512 KB, L3: 16-way shared 4 MB2.7 GHz
ARMv8.2-AM4 ("Cheetah")[96]2 cores, AArch64, 6-decode, 6-issue, 6-wide. superscalar, out-of-order64 KB I-cache / 64 KB D-cache, L2: 8-way private 1 MB, L3: 16-way shared 3 MB2.73 GHz
ARMv8.2-AM5 ("Lion")2 cores, AArch64, 6-decode, 6-issue, 6-wide. superscalar, out-of-order64 KB I-cache / 64 KB D-cache, L2: 8-way shared 2 MB, L3: 12-way shared 3 MB2.73 GHz

Timeline[edit]

The following table lists each core by the year it was announced.[97][98]

YearClassic coresCortex coresNeoverse cores
ARM1-6ARM7ARM8ARM9ARM10ARM11MicrocontrollerReal-timeApplication
(32-bit)
Application
(64-bit)
Application
(64-bit)
1985ARM1
1986ARM2
1989ARM3
1992ARM250
1993ARM60
ARM610
ARM700
1994ARM710
ARM7DI
ARM7TDMI
1995ARM710a
1996ARM810
1997ARM710T
ARM720T
ARM740T
1998ARM9TDMI
ARM940T
1999ARM9E-S
ARM966E-S
2000ARM920T
ARM922T
ARM946E-S
ARM1020T
2001ARM7TDMI-S
ARM7EJ-S
ARM9EJ-S
ARM926EJ-S
ARM1020E
ARM1022E
2002ARM1026EJ-SARM1136J(F)-S
2003ARM968E-SARM1156T2(F)-S
ARM1176JZ(F)-S
2004Cortex-M3
2005ARM11MPCoreCortex-A8
2006ARM996HS
2007Cortex-M1Cortex-A9
2008
2009Cortex-M0Cortex-A5
2010Cortex-M4(F)Cortex-A15
2011Cortex-R4
Cortex-R5
Cortex-R7
Cortex-A7
2012Cortex-M0+Cortex-A53
Cortex-A57
2013Cortex-A12
2014Cortex-M7(F)Cortex-A17
2015Cortex-A35
Cortex-A72
2016Cortex-M23
Cortex-M33(F)
Cortex-R8
Cortex-R52
Cortex-A32Cortex-A73
2017Cortex-A55
Cortex-A75
2018Cortex-M35P(F)Cortex-A65AE
Cortex-A76
Cortex-A76AE
2019Cortex-A77Neoverse E1
Neoverse N1
2020Cortex-M55(F)Cortex-R82Cortex-A78
Cortex-X1[99]
Neoverse V1[100]
2021Cortex-A510
Cortex-A710
Cortex-X2
Neoverse N2
2022Cortex-M85(F)Cortex-R52+Cortex-A715
Cortex-X3
2023Cortex-M52(F)Cortex-A520
Cortex-A720
Cortex-X4
2024Cortex-R82AECortex-A520AE
Cortex-A720AE
Neoverse N3
Neoverse V3
Neoverse V3AE

本文来自互联网用户投稿,该文观点仅代表作者本人,不代表本站立场。本站仅提供信息存储空间服务,不拥有所有权,不承担相关法律责任。如若转载,请注明出处:http://www.rhkb.cn/news/313502.html

如若内容造成侵权/违法违规/事实不符,请联系长河编程网进行投诉反馈email:809451989@qq.com,一经查实,立即删除!

相关文章

无人零售与传统便利店的竞争优势

无人零售与传统便利店的竞争优势 成本控制 • 无人零售 显著降低了人力成本,无需支付店员薪资和相关福利,且通过智能化管理减少能源消耗与维护费用,尤其在高租金和高人流区域效益突出。 • 传统便利店 则承担较高的人员开支,…

Nginx 四层和七层代理区别、配置

四层:通过报文中的目标地址和端口,加上负载均衡设备设置的服务器选择方式,决定最终选择的内部服务器,使用tcp、udp协议。 七层:"内容交换",通过报文中真正有意义的应用层内容,加上负…

GPT的全面历史和演变:从GPT-1到GPT-4

人工智能新篇章:GPT-4与人类互动的未来! 本文探讨了生成式预训练 Transformer (GPT) 的显着演变,提供了从开创性的 GPT-1 到复杂的 GPT-4 的旅程。 每次迭代都标志着重大的技术飞跃,深刻影响人工智能领域以及我们与技术的互动。 我…

MIS微调SAM模型实时交互UI界面

前言 SAM模型的基本介绍可见SAM(Segment Anything Model)大模型使用--point prompt_sam大模型-CSDN博客 针对Meta团队去年发布的SAM大模型在医学图像分割领域表现性能较差的情况,笔者收集了一些MIS领域的数据集对SAM的架构进行fine tune&am…

akSmart大带宽服务器基础配置科普

在数字化时代,服务器的性能和网络带宽成为业务发展的关键因素。RakSmart作为知名的服务器提供商,其大带宽服务器备受用户青睐。那么,RakSmart大带宽服务器的基础配置究竟有哪些呢?本文将为您揭开这一神秘面纱。 首先,我们来看看R…

信创需求激增,国产服务器操作系统赋能数字化转型

信创,即信息技术应用创新,是指在关键领域和环节推进信息技术的自主创新,构建安全可控的信息技术体系。随着数字化转型的加速,信创需求激增,国产服务器操作系统在其中扮演着至关重要的角色。国产服务器操作系统如何赋能…

爬虫入门——Request请求

目录 前言 一、Requests是什么? 二、使用步骤 1.引入库 2.请求 3.响应 三.总结 前言 上一篇爬虫我们已经提及到了urllib库的使用,为了方便大家的使用过程,这里为大家介绍新的库来实现请求获取响应的库。 一、Requests是什么&#xff1…

发布 Chrome/Edge浏览器extension扩展到应用商店

Chrom Extension发布流程 创建和发布自定义 Chrome 应用和扩展程序:https://support.google.com/chrome/a/answer/2714278?hlzh-Hans 在 Chrome 应用商店中发布:https://developer.chrome.com/docs/webstore/publish?hlzh-cn 注册开发者帐号&#…

Java 实际项目开发之最少必要知识汇总

一、面向过程和面向对象 程序设计语言分为面向过程和面向对象 对于面向过程,强调的是过程,关心的是我去做 对于面向对象,强调的是对象(万物皆对象),更关心的是我该让谁去做,这里的谁指的就是对象…

Hadoop+Spark大数据技术(微课版)曾国荪、曹洁版思维导图第四次作业 (第4章 HBase分布式DB)

1.简述Hbase的特点及与传统关系数据库的区别 HBase与传统关系数据库的区别 (1)数据类型 关系数据库具有丰富的数据类型,如字符串型、数值型、日期型、二进制型等。HBase只有字符串数据类型,数据的实际类型都是交由用户自己编写程序…

【Linux学习】Linux指令(四)

文章标题 🚀zip/unzip指令:🚀tar指令(重要):🚀uname –r指令:🚀关机指令🚀几个常用操作 🚀zip/unzip指令: zip 与 unzip的安装 yum i…

【Git学习笔记(二)】Git的基本操作

Git的基本操作 前言正文1、Git本地仓库创建和一些基本概念1.1 本地仓库创建1.2 本地项目创建1.3 仓库区和工作区1.4 暂存区1.5 提交1.6 查看状态1.7 查看日志 2、Git服务器和远程仓库创建2.1 搭建自己的Git服务器2.2 免费服务器2.2.1 github.com2.2.2 gitee.com2.2.3 创建远程仓…

MySQL基础知识——MySQL事务

事务背景 什么是事务? 一组由一个或多个数据库操作组成的操作组,能够原子的执行,且事务间相互独立; 简单来说,事务就是要保证一组数据库操作,要么全部成功,要么全部失败。 注:MyS…

TCGAplot在线版:输入基因,一键绘制TCGA基因表达与TMB相关性雷达图

1.TCGAplot简介 华科同济医院的王雄老师课题组利用TCGA数据库,开发了一个TCGA多组学数据泛癌分析和可视化R包TCGAplot[1]。用于泛癌表达以及基因表达与 TMB、MSI、TIME 和启动子甲基化之间相关性等分析。 2.TMB简介 肿瘤突变负荷(Tumor Mutation Burd…

【LLM】认识LLM

文章目录 1.LLM1.1 LLM简介1.2 LLM发展1.3 市面常见的LLM1.4 LLM涌现的能力 2.RAG2.1 RAG简介2.2 RAG 的工作流程2.3 RAG 和 Finetune 对比2.4 RAG的使用场景分析 3. LangChain3.1 LangChain简介3.2 LangChain的核心组件3.3 LangChain 入门 4.开发 RAG 应用的整体流程5. 环境配…

线上扭蛋机小程序成为年轻人创业选择

随着大众娱乐消费观的提升,带给消费者神秘感和惊喜感的扭蛋机开始走红,成为一个拥有广阔发展空间的行业。在当下二次元文化的火热下,扭蛋机的受众群体也遍布到了各个年龄层,深受大众的喜爱。 然而,玩家对扭蛋机的需求…

陇剑杯 省赛 攻击者1 CTF wireshark 流量分析

陇剑杯 省赛 攻击者1 题目 链接:https://pan.baidu.com/s/1KSSXOVNPC5hu_Mf60uKM2A?pwdhaek 提取码:haek ├───LogAnalize │ ├───linux简单日志分析 │ │ linux-log_2.zip │ │ │ ├───misc日志分析 │ │ acce…

伦敦银过夜费之——不能忽略的成本因素

众所周知,伦敦银是收益很好的投资品种,因为它采用了资金杠杆的原理,使投资者可以放大自己本金,而且还可以双向交易,并实现t0的资金回转。但在实际交易的过程中,投资者要充分考虑各种交易费用,对…

Zynq学习笔记--AXI 总线仿真(AXI VIP)

目录 1. 概述 2. Simulation with AXI VIP 2.1 axi_vip_pkg 2.2 design_1_axi_vip_0_0_pkg 2.3 参数指定 3. 实例化注意事项 3.1 名称对应关系 3.2 寄存器地址 3.3 block design 1. 概述 AXI Verification IP(AXI VIP)是一种用于验证AXI总线协议的…

MySQL行格式(row format)

MySQL行格式(row format) 表的行格式决定了其行的物理存储方式,这反过来又会影响查询和 DML 操作的性能。随着单个磁盘页面容纳更多行,查询和索引查找可以更快地工作,缓冲池中需要的高速缓存内存更少,写出…