Excavator (microarchitecture)

Microarchitecture by AMD From Wikipedia, the free encyclopedia

The Excavator Family 15h was an x86-64 CPU microarchitecture developed by AMD to for use in the company's CPU and APU CPUs, with commercial sales beginning in 2015. Excavator is the fourth and last microarchitecture derived from Bulldozer, and shares the same 28 nm manufacturing process with its predecessor, the Steamroller.[3][4] It was succeeded by the Zen microarchitecture in early 2017.[5][6]

LaunchedJune 2, 2015; 11 years ago (June 2, 2015)[1]
Common manufacturer
Sockets
Technologynode28 nm bulk silicon (GF28A)[2]
Quick facts General information, Launched ...
Excavator – Family 15h (4th-gen)
General information
LaunchedJune 2, 2015; 11 years ago (June 2, 2015)[1]
Common manufacturer
Physical specifications
Sockets
Architecture and classification
Technology node28 nm bulk silicon (GF28A)[2]
Instruction setAMD64 (x86-64)
Extensions
Products, models, variants
Core names
  • Carrizo
  • Bristol Ridge
  • Stoney Ridge
History
PredecessorSteamroller – Family 15h (3rd-gen)
SuccessorZen
Support status
iGPU unsupported
Close

The Excavator-based Carrizo chip is HSA 1.0 compliant.[7][8][9] An Excavator-based APU and CPU variant named Toronto for server and enterprise markets was also produced.[10]

Architecture

Excavator added hardware support for new instructions such as AVX2, BMI2 and RDRAND.[11] Excavator is designed using High Density (aka "Thin") Libraries normally used for GPUs to reduce electric energy consumption and die size, delivering a 30 percent increase in efficient energy use.[12] Excavator can process up to 15% more instructions per clock compared to AMD's previous core Steamroller.[13]

AMD's Fusion Controller Hub has been discontinued since the release of the Carrizo series of CPUs as it has been integrated into the same die as the rest of the CPU.[14]

Products

The following table shows features of AMD's processors with 3D graphics, including APUs (see also: List of AMD processors with 3D graphics).

More information Platform, High, standard and low power ...
Platform High, standard and low power Low and ultra-low power
CodenameServer Basic Toronto
Micro Kyoto
Desktop Performance Raphael Phoenix
Mainstream Llano Trinity Richland Kaveri Kaveri Refresh (Godavari) Carrizo Bristol Ridge Raven Ridge Picasso Renoir Cezanne
Entry
Basic Kabini Dalí
MobilePerformance Renoir Cezanne Rembrandt Dragon Range
Mainstream Llano Trinity Richland Kaveri Carrizo Bristol Ridge Raven Ridge Picasso Renoir
Lucienne
Cezanne
Barceló
Phoenix
Entry Dalí Mendocino
Basic Desna, Ontario, Zacate Kabini, Temash Beema, Mullins Carrizo-L Stoney Ridge Pollock
Embedded Trinity Bald Eagle Merlin Falcon,
Brown Falcon
Great Horned Owl Grey Hawk Ontario, Zacate Kabini Steppe Eagle, Crowned Eagle,
LX-Family
Prairie Falcon Banded Kestrel River Hawk
ReleasedAug 2011Oct 2012Jun 2013Jan 2014 2015Jun 2015Jun 2016Oct 2017Jan 2019Mar 2020 Jan 2021Jan 2022Sep 2022Jan 2023Jan 2011May 2013Apr 2014May 2015Feb 2016Apr 2019Jul 2020Jun 2022Nov 2022
CPU microarchitecture K10 Piledriver Steamroller Excavator "Excavator+"[15] Zen Zen+ Zen 2 Zen 3 Zen 3+ Zen 4 Bobcat Jaguar Puma Puma+[16] "Excavator+" Zen Zen+ "Zen 2+"
ISAx86-64 v1x86-64 v2x86-64 v3x86-64 v4x86-64 v1x86-64 v2x86-64 v3
Socket Desktop Performance N/a AM5 N/a N/a
Mainstream N/a AM4 N/a N/a
Entry FM1 FM2 FM2+ FM2+[a], AM4 AM4 N/a
Basic N/a N/a AM1 N/a FP5 N/a
Other FS1 FS1+, FP2 FP3 FP4 FP5 FP6 FP7 FL1 FP7
FP7r2
FP8
FT1 FT3 FT3b FP4 FP5 FT5 FP5 FT6
PCI Express version 2.0 3.0 4.0 5.0 4.0 2.0 3.0
CXLN/a N/a
Fab. (nm) GF 32SHP
(HKMG SOI)
GF 28SHP
(HKMG bulk)
GF 14LPP
(FinFET bulk)
GF 12LP
(FinFET bulk)
TSMC N7
(FinFET bulk)
TSMC N6
(FinFET bulk)
CCD: TSMC N5
(FinFET bulk)

cIOD: TSMC N6
(FinFET bulk)
TSMC 4nm
(FinFET bulk)
TSMC N40
(bulk)
TSMC N28
(HKMG bulk)
GF 28SHP
(HKMG bulk)
GF 14LPP
(FinFET bulk)
GF 12LP
(FinFET bulk)
TSMC N6
(FinFET bulk)
Die area (mm2)228246245245250210[17]156 180210CCD: (2x) 70
cIOD: 122
17875 (+ 28 FCH)107?125149~100
Min TDP (W)351712101565354.543.95106128
Max APU TDP (W)10095654517054182565415
Max stock APU base clock (GHz)33.84.14.13.73.83.63.73.84.03.34.74.31.752.222.23.22.61.23.352.8
Max APUs per node[b]11
Max core dies per CPU1211
Max CCX per core die1211
Max cores per CCX482424
Max CPU[c] cores per APU481682424
Max threads per CPU core1212
Integer pipeline structure3+32+24+24+2+11+3+3+1+21+1+1+12+24+24+2+1
i386, i486, i586, CMOV, NOPL, i686, PAE, NX bit, CMPXCHG16B, AMD-V, RVI, ABM, and 64-bit LAHF/SAHFYes Yes
IOMMU[d]N/av2v1v2
BMI1, AES-NI, CLMUL, and F16C YesN/aYes
MOVBEN/aYes
AVIC, BMI2, RDRAND, and MWAITX/MONITORX N/aYes
SME[e], TSME[e], ADX, SHA, RDSEED, SMAP, SMEP, XSAVEC, XSAVES, XRSTORS, CLFLUSHOPT, CLZERO, and PTE CoalescingN/aYes N/aYes
GMET, WBNOINVD, CLWB, QOS, PQE-BW, RDPID, RDPRU, and MCOMMITN/aYes N/aYes
MPK, VAESN/aYes N/a
SGXN/aN/a
FPUs per core10.5110.51
Pipes per FPU22
FPU pipe width128-bit256-bit80-bit128-bit256-bit
CPU instruction set SIMD levelSSE4a[f]AVX AVX2AVX-512SSSE3AVXAVX2
3DNow!3DNow!+N/a N/a
PREFETCH/PREFETCHWYes Yes
GFNIN/aYes N/a
AMXN/a
FMA4, LWP, TBM, and XOPN/aYesN/a N/aYesN/a
FMA3Yes Yes
AMD XDNAN/aYes N/a
L1 data cache per core (KiB)64163232
L1 data cache associativity (ways)2488
L1 instruction caches per core10.51 10.51
Max APU total L1 instruction cache (KiB)256128192256512256 64128 96 128
L1 instruction cache associativity (ways)2348 2 3 4 8
L2 caches per core10.5110.51
Max APU total L2 cache (MiB)424161212
L2 cache associativity (ways)168168
Max on-die L3 cache per CCX (MiB)N/a41632N/a4
Max 3D V-Cache per CCD (MiB)N/a64N/aN/a
Max total in-CCD L3 cache per APU (MiB)4816644
Max. total 3D V-Cache per APU (MiB)N/a64N/aN/a
Max. board L3 cache per APU (MiB)N/aN/a
Max total L3 cache per APU (MiB)48161284
APU L3 cache associativity (ways)1616
L3 cache schemeVictimVictim
Max. L4 cacheN/aN/a
Max stock DRAM supportDDR3-1866DDR3-2133DDR3-2133, DDR4-2400DDR4-2400DDR4-2933DDR4-3200, LPDDR4-4266DDR5-4800, LPDDR5-6400DDR5-5200DDR5-5600, LPDDR5x-7500DDR3L-1333DDR3L-1600DDR3L-1866DDR3-1866, DDR4-2400DDR4-2400DDR4-1600DDR4-3200LPDDR5-5500
Max DRAM channels per APU21212
Max stock DRAM bandwidth (GB/s) per APU29.86634.13238.40046.93268.256102.40083.200120.000 10.66612.80014.93319.20038.40012.80051.20088.000
GPU microarchitectureTeraScale 2 (VLIW5)TeraScale 3 (VLIW4)GCN 2nd genGCN 3rd genGCN 5th gen[18]RDNA 2RDNA 3TeraScale 2 (VLIW5)GCN 2nd genGCN 3rd gen[18]GCN 5th genRDNA 2
GPU instruction setTeraScale instruction setGCN instruction setRDNA instruction setTeraScale instruction setGCN instruction setRDNA instruction set
Max stock GPU base clock (MHz)60080084486611081250140021002400400 538600?847900120060013001900
Max stock GPU base GFLOPS[g]480614.4648.1886.71134.517601971.22150.43686.4102.4 86???345.6460.8230.41331.2486.4
3D engine[h]Up to 400:20:8Up to 384:24:6Up to 512:32:8Up to 704:44:16[19]Up to 512:32:8768:48:8128:8:480:8:4128:8:4Up to 192:12:8Up to 192:12:4192:12:4Up to 512:?:?128:?:?
IOMMUv1IOMMUv2IOMMUv1?IOMMUv2
Video decoderUVD 3.0UVD 4.2UVD 6.0VCN 1.0[20]VCN 2.1[21] VCN 2.2[21]VCN 3.1?UVD 3.0UVD 4.0UVD 4.2UVD 6.2VCN 1.0VCN 3.1
Video encoderN/aVCE 1.0VCE 2.0VCE 3.1N/aVCE 2.0VCE 3.4
AMD Fluid Motion No Yes No No Yes No
GPU power savingPowerPlayPowerTunePowerPlayPowerTune[22]
TrueAudioN/aYes[23]? N/aYes
FreeSync1
2
1
2
HDCP[i]?1.42.22.3?1.42.22.3
PlayReady[i]N/a3.0 not yetN/a3.0 not yet
Supported displays[j]2–32–433 (desktop)
4 (mobile, embedded)
42344
/drm/radeon[k][25][26]YesN/a YesN/a
/drm/amdgpu[k][27]N/aYes[28] N/aYes[28]
Close
  1. For FM2+ Excavator models: A8-7680, A6-7480 & Athlon X4 845.
  2. A PC would be one node.
  3. An APU combines a CPU and a GPU. Both have cores.
  4. Requires firmware support.
  5. Requires firmware support.
  6. No SSE4. No SSSE3.
  7. Single-precision performance is calculated from the base (or boost) core clock speed based on a FMA operation.
  8. To play protected video content, it also requires card, operating system, driver, and application support. A compatible HDCP display is also needed for this. HDCP is mandatory for the output of certain audio formats, placing additional constraints on the multimedia setup.
  9. To feed more than two displays, the additional panels must have native DisplayPort support.[24] Alternatively active DisplayPort-to-DVI/HDMI/VGA adapters can be employed.
  10. DRM (Direct Rendering Manager) is a component of the Linux kernel. Support in this table refers to the most current version.

Processors

APU lines

There are three APU lines announced or released:

  1. Budget and mainstream markets (desktop and mobile): Carrizo APU
    • The Carrizo mobile APUs were launched in 2015 based on Excavator x86 cores and featuring Heterogeneous System Architecture for integrated task sharing between CPUs and GPUs, which allows a GPU to perform compute functions, which is claimed provide greater performance increases than shrinking the feature size alone.[7]
    • Carrizo desktop APUs were launched in 2018. The mainstream product (A8-7680) has 4 Excavator cores and a GPU based on GCN1.2 architecture. Also, an entry-level APU (A6-7480) with 2 Excavator cores is also launched.
  2. Budget and mainstream markets (desktop and mobile): Bristol Ridge, and Stoney Ridge (for entry level notebooks), APUs[29]
    • Bristol Ridge APUs utilize socket AM4 and DDR4 RAM
    • Bristol Ridge APUs have up to 4 Excavator CPU cores and up to 8 3rd generation GCN GPU cores
    • Up to a 20% CPU performance increase over Carrizo
    • TDP of 15W to 65W, 15–35W for mobile
  3. Enterprise and server markets: Toronto APU
    • The Toronto APU for server and enterprise markets featured four x86 Excavator CPU core modules and Volcanic Islands integrated GPU core.
    • The Excavator cores has a greater advantage with IPC than Steamroller. The improvement is 4–15%.
    • Support for HSA/hUMA, DDR3/DDR4, PCIe 3.0, GCN 1.2[7][10][6]
    • The Toronto APU was available in BGA and SoC variants. The SoC variant had the southbridge on the same die as the APU to save space and power and to optimize workloads.
    • A complete system with a Toronto APU would have a maximum power usage of 70 W.[10]

CPU Desktop lines

There are no CPUs built on Steamroller (3rd gen Bulldozer) or Excavator (4th gen Bulldozer) architectures on high-end desktop platforms.

Excavator CPU for Desktop announced on February 2, 2016, named Athlon X4 845.[30] In 2017, three more desktop CPUs (Athlon X4 9x0) were launched. They come in Socket AM4, with a TDP of 65W. In fact, they are APUs with their graphics cores disabled.

More information Brand Name, Model Number ...
List of desktop Excavator CPUs
Brand

Name

Model

Number

Code

Name

Freq. (GHz) Cores TDP

(W)

Socket Cache PCI Express 3.0 Relative IPC Locked
Base TurboL1DL2
Athlon X4 845 Carrizo3.5 3.8465 FM2+4x

32KB

2x

1MB

x81.0 Yes
940 Bristol Ridge3.2 3.6 AM4x161.1 No
950 3.5 3.8
970 3.8 4.0
Close

Server lines

The AMD Opteron roadmaps for 2015 show the Excavator-based Toronto APU and Toronto CPU intended for 1 Processor (1P) cluster applications:[10]

  • For 1P Web and Enterprise Services Clusters:
    • Toronto CPU – quad-core x86 Excavator architecture
    • plans for Cambridge CPU – 64-bit AArch64 core
  • For 1P Compute and Media Clusters:
    • Toronto APU – quad-core x86 Excavator architecture
  • For 2P/4P Servers:

References

Related Articles

Wikiwand AI