imported>Comrade John：增加或調整內部連結

2025-04-18T15:19:04Z

增加或調整內部連結

新页面

{{Sticky table start}}
{{sticky header}}
{| class="wikitable sortable sticky-header-multi sticky-table-row1 sticky-table-col1" style="font-size: 80%; text-align: center;"
|-
! rowspan=2 style="width:16em" | 型号
! rowspan=2 | [[微架構]]
! rowspan=2 | {{nowrap|发布时间}}
! rowspan=2 | 芯片
! rowspan=2 | 核心 频率
! colspan=3 | 渲染器
! colspan=5 | 内存
! colspan=3 | 算力{{efn|name="Calculate"|group="NvidiaTesla"|To calculate the processing power see {{link-en|Tesla (微架构)}}, [[費米微架構]], [[Kepler (微架构)]], {{link-en|Maxwell (microarchitecture)}}, or [[帕斯卡 (微架构)]]. A number range specifies the minimum and maximum processing power at, respectively, the base clock and maximum boost clock.}}
! rowspan=2 | CUDA 运算能力{{efn|group="NvidiaTesla|Core architecture version according to the [[CUDA]] programming guide.}}
! rowspan=2 | [[热设计功耗]]
! rowspan=2 | 注释/ 外形尺寸
|-
! {{nowrap|CUDA核心}} （总数）
! 基础 频率
! 最大 增强 频率{{efn|name="GPUBoost"|group="NvidiaTesla"|GPU Boost is a default feature that increases the core clock rate while remaining under the card's predetermined power budget. Multiple boost clocks are available, but this table lists the highest clock supported by each card.<ref>{{cite web |url=https://www.nvidia.com/content/PDF/kepler/nvidia-gpu-boost-tesla-k40-06767-001-v02.pdf |title=Nvidia GPU Boost For Tesla |date=January 2014 |access-date=7 December 2015 |archive-date=2020-05-16 |archive-url=https://web.archive.org/web/20200516090655/https://www.nvidia.com/content/PDF/kepler/nvidia-gpu-boost-tesla-k40-06767-001-v02.pdf |dead-url=no }}</ref>}}
! 总线 类型
! 总线字长
! 容量大小
! 频率
! 带宽
! [[半精度浮点数|半精度]] Tensor Core FP32 累加
! [[單精度浮點數|單精度]] （[[乘積累加運算|MAD]]或[[积和熔加运算|FMA]]）
! [[雙精度浮點數|雙精度]] （[[积和熔加运算|FMA]]）
|-
! 单位 !! !! !! !! [[赫兹|MHz]] !! !! MHz !! MHz !! !! [[位元|bit]] !! [[吉字节|GB]] !! {{link-en|每秒传输|Transfers per second|MT/s}} !! GB/s !! [[每秒浮點運算次數|GFLOPs]] !! GFLOPs !! GFLOPs !! !! [[瓦特|W]] !!
|-
! style="text-align:left;" | C870 GPU Computing Module{{efn|group="NvidiaTesla"|name="Assumed8800"|Specifications not specified by Nvidia assumed to be based on the [[NVIDIA GeForce 8]]GTX}}
| rowspan=9 | {{link-en|Tesla (微架构)|Tesla (microarchitecture)|Tesla}}
| 2007年 5月2日
| 1× G80
| 600
| 128
| 1350
| {{n/a}}
| GDDR3
| 384
| 1.5
| 1600
| 76.8
| {{no}}
| 345.6
| {{no}}
| 1.0
| 170.9
| Internal PCIe GPU (full-height, dual-slot)
|-
! style="text-align:left;" | D870 Deskside Computer{{efn|group="NvidiaTesla"|name="Assumed8800"}}
| 2007年 5月2日
| 2× G80
| 600
| 256
| 1350
| {{n/a}}
| GDDR3
| 2× 384
| 2× 1.5
| 1600
| 2× 76.8
| {{no}}
| 691.2
| {{no}}
| 1.0
| 520
| Deskside or 3U {{link-en|19-inch rack}} external GPUs
|-
! style="text-align:left;" | S870 GPU Computing Server{{efn|group="NvidiaTesla"|name="Assumed8800"}}
| 2007年 5月2日
| 4× G80
| 600
| 512
| 1350
| {{n/a}}
| GDDR3
| 4× 384
| 4× 1.5
| 1600
| 4× 76.8
| {{no}}
| 1382.4
| {{no}}
| 1.0
|
| 1U {{link-en|19-inch rack}} external GPUs, connect via 2× PCIe (×16)
|-
! style="text-align:left;" | C1060 GPU Computing Module{{efn|name="Assumed280"|group="NvidiaTesla"|Specifications not specified by Nvidia assumed to be based on the [[NVIDIA GeForce 200]]}}
| 2009年 4月9日
| 1× GT200
| 602
| 240
| 1296<ref>{{cite web |url=http://www.nvidia.com/docs/IO/56483/Tesla_C1060_boardSpec_v03.pdf |format=PDF |title=Tesla C1060 Computing Processor Board |website=Nvidia.com |accessdate=2015-12-11 |archive-date=2020-11-24 |archive-url=https://web.archive.org/web/20201124222334/https://www.nvidia.com/docs/IO/56483/Tesla_C1060_boardSpec_v03.pdf |dead-url=no }}</ref>
| {{n/a}}
| GDDR3
| 512
| 4
| 1600
| 102.4
| {{no}}
| 622.08
| 77.76
| 1.3
| 187.8
| Internal PCIe GPU (full-height, dual-slot)
|-
! style="text-align:left;" | S1070 GPU Computing Server "400 configuration"{{efn|name="Assumed280"|group="NvidiaTesla"}}
| 2008年 6月1日
|rowspan=2| 4× GT200
|rowspan=2| 602
|rowspan=2| 960
| 1296
| {{n/a}}
|rowspan=2| GDDR3
|rowspan=2| 4× 512
|rowspan=2| 4× 4
|rowspan=2| 1538.4
|rowspan=2| 4× 98.5
| {{no}}
| 2488.3
| 311.0
|rowspan=2| 1.3
|rowspan=2| 800
|rowspan=2| 1U {{link-en|19-inch rack}} external GPUs, connect via 2× PCIe (×8 or ×16)
|-
! style="text-align:left;" | S1070 GPU Computing Server "500 configuration"{{efn|name="Assumed280"|group="NvidiaTesla"}}
|
| 1440
| {{n/a}}
| {{no}}
| 2764.8
| 345.6
|-
! style="text-align:left;" | S1075 GPU Computing Server{{efn|name="Assumed280"|group="NvidiaTesla"}}<ref>{{cite web |url=http://forums.nvidia.com/index.php?showtopic=80850 |title=Difference between Tesla S1070 and S1075 |date=31 October 2008 |access-date=January 29, 2017 |quote=S1075 has one interface card |archive-date=2012-02-26 |archive-url=https://web.archive.org/web/20120226001759/http://forums.nvidia.com/index.php?showtopic=80850 |dead-url=no }}</ref>
| 2008年 6月1日
| 4× GT200
| 602
| 960
| 1440
| {{n/a}}
| GDDR3
| 4× 512
| 4× 4
| 1538.4
| 4× 98.5
| {{no}}
| 2764.8
| 345.6
| 1.3
|
| 1U {{link-en|19-inch rack}} external GPUs, connect via 1× PCIe (×8 or ×16)
|-
! style="text-align:left;" | Quadro Plex 2200 D2 Visual Computing System{{efn|group="NvidiaTesla"|name="Assumed5800"|Specifications not specified by Nvidia assumed to be based on the Quadro FX 5800}}
| 2008年 7月25日
| 2× GT200GL
| 648
| 480
| 1296
| {{n/a}}
| GDDR3
| 2× 512
| 2× 4
| 1600
| 2× 102.4
| {{no}}
| 1244.2
| 155.5
| 1.3
|
| Deskside or 3U {{link-en|19-inch rack}} external GPUs with 4 dual-link DVI outputs
|-
! style="text-align:left;" | Quadro Plex 2200 S4 Visual Computing System{{efn|group="NvidiaTesla"|name="Assumed5800"}}
| 2008年 7月25日
| 4× GT200GL
| 648
| 960
| 1296
| {{n/a}}
| GDDR3
| 4× 512
| 4× 4
| 1600
| 4× 102.4
| {{no}}
| 2488.3
| 311.0
| 1.3
| 1200
| 1U {{link-en|19-inch rack}} external GPUs, connect via 2× PCIe (×8 or ×16)
|-
! style="text-align:left;" | C2050 GPU Computing Module<ref name="nvidia1">{{cite web |url=http://www.nvidia.com/docs/io/43395/tesla_c2050_board_specification.pdf |format=PDF |title=Tesla C2050 and Tesla C2070 Computing Processor |website=Nvidia.com |accessdate=2015-12-11 |archive-date=2020-11-24 |archive-url=https://web.archive.org/web/20201124215355/https://www.nvidia.com/docs/io/43395/tesla_c2050_board_specification.pdf |dead-url=no }}</ref>
| rowspan=8 | [[費米微架構|Fermi]]
| 2011年 7月25日
|rowspan=2| 1× GF100
|rowspan=2| 575
|rowspan=2| 448
|rowspan=2| 1150
| {{n/a}}
|rowspan=2| GDDR5
|rowspan=2| 384
|rowspan=2| 3{{efn|name="ECC"|group="NvidiaTesla|With ECC on, a portion of the dedicated memory is used for ECC bits, so the available user memory is reduced by 12.5%. (e.g. 4 GB total memory yields 3.5 GB of user available memory.)}}
| 3000
| 144
| {{no}}
|rowspan=2| 1030.4
|rowspan=2| 515.2
|rowspan=2| 2.0
| 247
|rowspan=2| Internal PCIe GPU (full-height, dual-slot)
|-
! style="text-align:left;" | M2050 GPU Computing Module<ref>{{cite web |url=http://www.nvidia.com/docs/io/43395/bd-05238-001_v03.pdf |format=PDF |title=Tesla M2050 and Tesla M2070/M2070Q Dual-Slot Computing Processor Modules |website=Nvidia.com |accessdate=2015-12-11 |archive-date=2020-11-24 |archive-url=https://web.archive.org/web/20201124215116/https://www.nvidia.com/docs/io/43395/bd-05238-001_v03.pdf |dead-url=no }}</ref>
| 2011年 7月25日
| {{n/a}}
| 3092
| 148.4
| {{no}}
| 225
|-
! style="text-align:left;" | C2070 GPU Computing Module<ref name="nvidia1"/>
| July 25, 2011
|rowspan=3| 1× GF100
|rowspan=3| 575
|rowspan=3| 448
|rowspan=3| 1150
| {{n/a}}
|rowspan=3| GDDR5
|rowspan=3| 384
|rowspan=3| 6{{efn|name="ECC"|group="NvidiaTesla"}}
| 3000
| 144
| {{no}}
|rowspan=3| 1030.4
|rowspan=3| 515.2
|rowspan=3| 2.0
| 247
|rowspan=3| Internal PCIe GPU (full-height, dual-slot)
|-
! style="text-align:left;" | C2075 GPU Computing Module<ref>{{cite web |url=https://www.nvidia.com/docs/IO/43395/BD-05880-001_v02.pdf |format=PDF |title=Tesla C2075 Computing Processor Board |website=Nvidia.com |accessdate=2015-12-11 |archive-date=2020-11-24 |archive-url=https://web.archive.org/web/20201124212405/https://www.nvidia.com/docs/IO/43395/BD-05880-001_v02.pdf |dead-url=no }}</ref>
| 2011年 7月25日
| {{n/a}}
| 3000
| 144
| {{no}}
| 225
|-
! style="text-align:left;" | M2070/M2070Q GPU Computing Module<ref>{{cite web |last=Hand |first=Randall |url=http://www.vizworld.com/2010/08/nvidia-tesla-m2050-m2070m2070q-specs-online/ |title=NVidia Tesla M2050 & M2070/M2070Q Specs OnlineVizWorld.com |website=VizWorld.com |date=2010-08-23 |accessdate=2015-12-11 |archive-date=2020-08-17 |archive-url=https://web.archive.org/web/20200817185337/https://vizworld.com/2010/08/nvidia-tesla-m2050-m2070m2070q-specs-online/ |dead-url=no }}</ref>
| 2011年 7月25日
| {{n/a}}
| 3132
| 150.336
| {{no}}
| 225
|-
! style="text-align:left;" | M2090 GPU Computing Module<ref>{{cite web |url=http://www.nvidia.com/docs/IO/43395/Tesla-M2090-Board-Specification.pdf |format=PDF |title=Tesla M2090 Dual-Slot Computing Processor Module |website=Nvidia.com |accessdate=2015-12-11 |archive-date=2020-10-26 |archive-url=https://web.archive.org/web/20201026170151/https://www.nvidia.com/docs/IO/43395/Tesla-M2090-Board-Specification.pdf |dead-url=no }}</ref>
| 2011年 7月25日
| 1× GF110
| 650
| 512
| 1300
| {{n/a}}
| GDDR5
| 384
| 6{{efn|name="ECC"|group="NvidiaTesla"}}
| 3700
| 177.6
| {{no}}
| 1331.2
| 665.6
| 2.0
| 225
| Internal PCIe GPU (full-height, dual-slot)
|-
! style="text-align:left;" | S2050 GPU Computing Server
| 2011年 7月25日
|rowspan=2| 4× GF100
|rowspan=2| 575
|rowspan=2| 1792
|rowspan=2| 1150
| {{n/a}}
|rowspan=2| GDDR5
|rowspan=2| 4× 384
| 4× 3{{efn|name="ECC"|group="NvidiaTesla"}}
|rowspan=2| 3
|rowspan=2| 4× 148.4
| {{no}}
|rowspan=2| 4121.6
|rowspan=2| 2060.8
|rowspan=2| 2.0
|rowspan=2| 900
|rowspan=2| 1U {{link-en|19-inch rack}} external GPUs, connect via 2× PCIe (×8 or ×16)
|-
! style="text-align:left;" | S2070 GPU Computing Server
|
| {{n/a}}
| 4× 6{{efn|name="ECC"|group="NvidiaTesla"}}
| {{no}}
|-
! style="text-align:left;" | K10 GPU accelerator<ref>{{cite web |url=http://www.nvidia.com/content/PDF/kepler/Tesla_K10_BD-06280-001_v05.pdf |format=PDF |title=Tesla K10 GPU accelerator |website=Nvidia.com |accessdate=2015-12-11 |archive-date=2020-11-24 |archive-url=https://web.archive.org/web/20201124214105/https://www.nvidia.com/content/PDF/kepler/Tesla_K10_BD-06280-001_v05.pdf |dead-url=no }}</ref>
| rowspan=5 | [[开普勒微架构|Kepler]]
| 2012年 5月1日
| 2× GK104
| {{n/a}}
| 3072
| 745
| {{dunno}}
| GDDR5
| 2× 256
| 2× 4
| 5000
| 2× 160
| {{no}}
| 4577
| 190.7
| 3.0
| 225
| Internal PCIe GPU (full-height, dual-slot)
|-
! style="text-align:left;" | K20 GPU accelerator<ref>{{cite web |url=http://www.nvidia.com/content/PDF/kepler/Tesla-K20-Active-BD-06499-001-v02.pdf |format=PDF |title=Tesla K20 GPU active accelerator |website=Nvidia.com |accessdate=2015-12-11 |archive-date=2017-07-12 |archive-url=https://web.archive.org/web/20170712234519/http://www.nvidia.com/content/PDF/kepler/Tesla-K20-Active-BD-06499-001-v02.pdf |dead-url=no }}</ref><ref>{{cite web |url=https://www.nvidia.com/content/PDF/kepler/Tesla-K20-Passive-BD-06455-001-v05.pdf |format=PDF |title=Tesla K20 GPU accelerator |website=Nvidia.com |accessdate=2015-12-11 |archive-date=2020-03-26 |archive-url=https://web.archive.org/web/20200326145257/https://www.nvidia.com/content/PDF/kepler/Tesla-K20-Passive-BD-06455-001-v05.pdf |dead-url=no }}</ref>
| 2012年 11月12日
| 1× GK110
| {{n/a}}
| 2496
| 706
| 758
| GDDR5
| 320
| 5
| 5200
| 208
| {{no}}
| 3524
| 1175
| 3.5
| 225
| Internal PCIe GPU (full-height, dual-slot)
|-
! style="text-align:left;" | K20X GPU accelerator<ref>{{cite web |url=http://www.nvidia.com/content/pdf/kepler/tesla-k20x-bd-06397-001-v07.pdf |format=PDF |title=Tesla K20X GPU accelerator |website=Nvidia.com |accessdate=2015-12-11 |archive-date=2020-11-24 |archive-url=https://web.archive.org/web/20201124204809/https://www.nvidia.com/content/pdf/kepler/tesla-k20x-bd-06397-001-v07.pdf |dead-url=no }}</ref>
| 2012年 11月12日
| 1× GK110
| {{n/a}}
| 2688
| 732
| {{dunno}}
| GDDR5
| 384
| 6
| 5200
| 250
| {{no}}
| 3935
| 1312
| 3.5
| 235
| Internal PCIe GPU (full-height, dual-slot)
|-
! style="text-align:left;" | K40 GPU accelerator<ref>{{cite web |url=http://www.nvidia.com/content/PDF/kepler/Tesla-K40-PCIe-Passive-Board-Spec-BD-06902-001_v05.pdf |format=PDF |title=Tesla K40 GPU accelerator |website=Nvidia.com |accessdate=2015-12-11 |archive-date=2020-06-17 |archive-url=https://web.archive.org/web/20200617151953/https://www.nvidia.com/content/PDF/kepler/Tesla-K40-PCIe-Passive-Board-Spec-BD-06902-001_v05.pdf |dead-url=no }}</ref>
| 2013年 10月8日
| 1× GK110B
| {{n/a}}
| 2880
| 745
| 875
| GDDR5
| 384
| 12{{efn|name="ECC"|group="NvidiaTesla"}}
| 6000
| 288
| {{no}}
| 4291–5040
| 1430–1680
| 3.5
| 235
| Internal PCIe GPU (full-height, dual-slot)
|-
! style="text-align:left;" | K80 GPU accelerator<ref>{{cite web |url=http://images.nvidia.com/content/pdf/kepler/Tesla-K80-BoardSpec-07317-001-v05.pdf |format=PDF |title=Tesla K80 GPU accelerator |website=Images.nvidia.com |accessdate=2015-12-11 |archive-date=2018-07-12 |archive-url=https://web.archive.org/web/20180712151016/http://images.nvidia.com/content/pdf/kepler/Tesla-K80-BoardSpec-07317-001-v05.pdf |dead-url=no }}</ref>
| 2014年 11月17日
| 2× GK210
| {{n/a}}
| 4992
| 560
| 875
| GDDR5
| 2× 384
| 2× 12
| 5000
| 2× 240
| {{no}}
| 5591–8736
| 1864–2912
| 3.7
| 300
| Internal PCIe GPU (full-height, dual-slot)
|-
! style="text-align:left;" | M4 GPU accelerator<ref name="anandtech1">{{cite web |url=http://www.anandtech.com/show/9776/nvidia-announces-tesla-m40-m4-server-cards-data-center-machine-learning |title=Nvidia Announces Tesla M40 & M4 Server Cards - Data Center Machine Learning |website=Anandtech.com |date= |accessdate=2015-12-11 |archive-date=2020-11-08 |archive-url=https://web.archive.org/web/20201108124231/https://www.anandtech.com/show/9776/nvidia-announces-tesla-m40-m4-server-cards-data-center-machine-learning |dead-url=no }}</ref><ref name="nvidia2">{{cite web |url=http://devblogs.nvidia.com/parallelforall/accelerating-hyperscale-datacenter-applications-tesla-gpus/ |title=Accelerating Hyperscale Datacenter Applications with Tesla GPUs | Parallel Forall |website=Devblogs.nvidia.com |date=2015-11-10 |accessdate=2015-12-11 |archive-date=2017-07-09 |archive-url=https://web.archive.org/web/20170709172435/https://devblogs.nvidia.com/parallelforall/accelerating-hyperscale-datacenter-applications-tesla-gpus/ |dead-url=no }}</ref>
| rowspan=5 | [[Maxwell微架构|Maxwell]]
| 2015年 11月10日
| 1× GM206
| {{n/a}}
| 1024
| 872
| 1072
| GDDR5
| 128
| 4
| 5500
| 88
| {{no}}
| 1786–2195
| 55.81–68.61
| 5.2
| 50–75
| Internal PCIe GPU (half-height, single-slot)
|-
! style="text-align:left;" | M6 GPU accelerator<ref>{{cite web |url=http://images.nvidia.com/content/pdf/tesla/tesla-m6-product-brief.pdf |format=PDF |title=Tesla M6 |website=Images.nvidia.com |accessdate=2016-05-28 |archive-date=2020-10-22 |archive-url=https://web.archive.org/web/20201022155840/http://images.nvidia.com/content/pdf/tesla/tesla-m6-product-brief.pdf |dead-url=no }}</ref>
| 2015年 8月30日
| 1× GM204-995-A1
| {{n/a}}
| 1536
| 722
| 1051
| GDDR5
| 256
| 8
| 4600
| 147.2
| {{no}}
| 2218–3229
| 69.3–100.9
| 5.2
| 75–100
| Internal MXM GPU
|-
! style="text-align:left;" | M10 GPU accelerator<ref>{{cite web |url=http://images.nvidia.com/content/pdf/tesla/Tesla-M10-Product-Brief.pdf |format=PDF |title=Tesla M10 |website=Images.nvidia.com |accessdate=2016-10-29 |archive-date=2017-05-10 |archive-url=https://web.archive.org/web/20170510101626/http://images.nvidia.com/content/pdf/tesla/Tesla-M10-Product-Brief.pdf |dead-url=no }}</ref>
|
| 4× GM107
| {{n/a}}
| 2560
| 1033
| {{dunno}}
| GDDR5
| 4× 128
| 4× 8
| 5188
| 4× 83
| {{no}}
| 5289
| 165.3
| 5.2
| 225
| Internal PCIe GPU (full-height, dual-slot)
|-
! style="text-align:left;" | M40 GPU accelerator<ref name="nvidia2"/><ref>{{cite web |url=http://images.nvidia.com/content/tesla/pdf/tesla-m40-product-brief.pdf |format=PDF |title=Tesla M40 |website=Images.nvidia.com |accessdate=2015-12-11 |archive-date=2016-10-21 |archive-url=https://web.archive.org/web/20161021031653/http://images.nvidia.com/content/tesla/pdf/tesla-m40-product-brief.pdf |dead-url=no }}</ref>
| 2015年 11月10日
| 1× GM200
| {{n/a}}
| 3072
| 948
| 1114
| GDDR5
| 384
| 12
| 6000
| 288
| {{no}}
| 5825–6844
| 182.0–213.9
| 5.2
| 250
| Internal PCIe GPU (full-height, dual-slot)
|-
! style="text-align:left;" | M60 GPU accelerator<ref>{{cite web |url=http://images.nvidia.com/content/pdf/tesla/tesla-m60-product-brief.pdf |format=PDF |title=Tesla M60 |website=Images.nvidia.com |accessdate=2016-05-27 |archive-date=2020-10-22 |archive-url=https://web.archive.org/web/20201022152228/http://images.nvidia.com/content/pdf/tesla/tesla-m60-product-brief.pdf |dead-url=no }}</ref>
| 2015年 8月30日
| 2× GM204-895-A1
| {{n/a}}
| 4096
| 899
| 1178
| GDDR5
| 2× 256
| 2× 8
| 5000
| 2× 160
| {{no}}
| 7365–9650
| 230.1–301.6
| 5.2
| 225–300
| Internal PCIe GPU (full-height, dual-slot)
|-
! style="text-align:left;" | P4 GPU accelerator<ref name="Anand_P40">{{cite news|last1=Smith|first1=Ryan|title=Nvidia Announces Tesla P40 & Tesla P4 - Network Inference, Big & Small|url=http://www.anandtech.com/show/10675/nvidia-announces-tesla-p40-tesla-p4|accessdate=13 September 2016|publisher=Anandtech|date=13 September 2016|archive-date=2021-01-16|archive-url=https://web.archive.org/web/20210116101008/https://www.anandtech.com/show/10675/nvidia-announces-tesla-p40-tesla-p4|dead-url=no}}</ref>
| rowspan=6 | [[帕斯卡 (微架构)|Pascal]]
| 2016年 9月13日
| 1× GP104
| {{n/a}}
| 2560
| 810
| 1063
| GDDR5
| 256
| 8
| 6000
| 192.0
| {{no}}
| 4147–5443
| 129.6–170.1
| 6.1
| 50-75
| [[PCI Express]] card
|-
|-
! style="text-align:left;" | P6 GPU accelerator<ref>{{cite web |url=https://www.nvidia.com/content/dam/en-zz/Solutions/design-visualization/solutions/resources/documents1/Tesla-P6-Product-Brief.pdf |format=PDF |title=Tesla P6 |website=www.nvidia.com |accessdate=2019-03-07 |archive-date=2020-07-28 |archive-url=https://web.archive.org/web/20200728182944/https://www.nvidia.com/content/dam/en-zz/Solutions/design-visualization/solutions/resources/documents1/Tesla-P6-Product-Brief.pdf |dead-url=no }}</ref><ref>{{cite web |url=https://www.techpowerup.com/gpu-specs/tesla-p6.c3036 |title=Tesla P6 Specs |website=www.techpowerup.com |accessdate=2019-03-07 |archive-date=2020-07-28 |archive-url=https://web.archive.org/web/20200728180652/https://www.techpowerup.com/gpu-specs/tesla-p6.c3036 |dead-url=no }}</ref>
| 2017年 3月24日
| 1× GP104-995-A1
| {{n/a}}
| 2048
| 1012
| 1506
| GDDR5
| 256
| 16
| 3003
| 192.2
| {{no}}
| 6169
| 192.8
| 6.1
| 90
| [[行動PCI Express模組]] card
|-
|-
! style="text-align:left;" | P40 GPU accelerator<ref name="Anand_P40"/>
| 2016年 9月13日
| 1× GP102
| {{n/a}}
| 3840
| 1303
| 1531
| GDDR5
| 384
| 24
| 7200
| 345.6
| {{no}}
| 10007–11758
| 312.7–367.4
| 6.1
| 250
| [[PCI Express]] card
|-
|-
! style="text-align:left;" | P100 GPU accelerator (mezzanine)<ref>{{cite news|last1=Smith|first1=Ryan|title=Nvidia Announces Tesla P100 Accelerator - Pascal GP100 for HPC|url=http://www.anandtech.com/show/10222/nvidia-announces-tesla-p100-accelerator-pascal-power-for-hpc|accessdate=5 April 2016|agency=Anandtech.com|publisher=Anandtech.com|date=5 April 2016|archive-date=2016-07-30|archive-url=https://web.archive.org/web/20160730213045/http://www.anandtech.com/show/10222/nvidia-announces-tesla-p100-accelerator-pascal-power-for-hpc|dead-url=no}}</ref><ref>{{cite news|last1=Harris|first1=Mark|title=Inside Pascal: Nvidia’s Newest Computing Platform|url=https://devblogs.nvidia.com/parallelforall/inside-pascal/|accessdate=13 September 2016|archive-date=2017-05-07|archive-url=https://web.archive.org/web/20170507110037/https://devblogs.nvidia.com/parallelforall/inside-pascal/|dead-url=no}}</ref>
| 2016年 4月5日
| 1× GP100-890-A1
| {{n/a}}
| rowspan=3 | 3584
| 1328
| 1480
| rowspan=3 | [[高頻寬記憶體|HBM2]]
| rowspan=2 | 4096
| rowspan=2 | 16
| rowspan=3 | 1430
| rowspan=2 | 732
| {{no}}
| 9519–10609
| 4760–5304
| rowspan=3 | 6.0
| 300
| [[NVLink]] card
|-
|-
! style="text-align:left;" | P100 GPU accelerator (16 GB card)<ref name="P100-PCIe">{{cite news|last1=Smith|first1=Ryan|title=NVidia Announces PCI Express Tesla P100|url=http://www.anandtech.com/show/10433/nvidia-announces-pci-express-tesla-p100|accessdate=21 June 2016|publisher=Anandtech.com|date=20 June 2016|archive-date=2020-11-11|archive-url=https://web.archive.org/web/20201111203959/https://www.anandtech.com/show/10433/nvidia-announces-pci-express-tesla-p100|dead-url=no}}</ref>
| 2016年 6月20日
| rowspan=2 | 1× GP100
| {{n/a}}
| rowspan=2 | 1126
| rowspan=2 | 1303
| {{no}}
| 8071‒9340
| 4036‒4670
| rowspan=2 | 250
| rowspan=2 | [[PCI Express]] card
|-
|-
! style="text-align:left;" | P100 GPU accelerator (12 GB card)<ref name="P100-PCIe"></ref>
| 2016年 6月20日
| {{n/a}}
| 3072
| 12
| 549
| {{no}}
| 8071‒9340
| 4036‒4670
|-
|-
! style="text-align:left;" | V100 GPU accelerator (mezzanine)<ref name='v100_1'>{{cite news|last1=Smith|first1=Ryan|title=The Nvidia GPU Technology Conference 2017 Keynote Live Blog|url=http://www.anandtech.com/show/11360/the-nvidia-gpu-tech-conference-2017-keynote-live-blog|access-date=10 May 2017|publisher=Anandtech|date=10 May 2017}}</ref><ref name='v100_2'>{{cite news|last1=Smith|first1=Ryan|title=NVIDIA Volta Unveiled: GV100 GPU and Tesla V100 Accelerator Announced|url=http://www.anandtech.com/show/11367/nvidia-volta-unveiled-gv100-gpu-and-tesla-v100-accelerator-announced|access-date=10 May 2017|publisher=Anandtech|date=10 May 2017}}</ref><ref name="V100_3">{{cite news|last1=Oh|first1=Nate|title=NVIDIA Formally Announces V100: Available later this Year|url=http://www.anandtech.com/show/11559/nvidia-formally-announces-pcie-tesla-v100-available-later-this-year|access-date=20 June 2017|publisher=Anandtech.com|date=20 June 2017}}</ref>
| rowspan=3| [[Volta (微架构)|Volta]]
| 2017年 5月10日
| 1× GV100-895-A1
| {{n/a}}
| rowspan=3| 5120
| {{unk}}
| 1,455
| rowspan=3|HBM2
| rowspan=3|4,096
| rowspan=2|16 or 32
| rowspan=2|1,750
| rowspan=2|900
| 119,192
| 14,899
| 7,450
| rowspan=3| 7.0
| 300
| [[SXM (socket)|SXM]] card
|-
|-
! style="text-align:left;" | V100 GPU accelerator (PCIe card)<ref name='v100_1' /><ref name='v100_2' /><ref name="V100_3" />
| 2017年 6月21日
| 1× GV100
| {{n/a}}
| {{unk}}
| 1,370
| 112,224
| 14,028
| 7,014
| 250
| PCIe card
|-
|-
! style="text-align:left;" | V100 GPU accelerator (PCIe FHHL card)
| 2018年 3月27日
| 1× GV100
| {{n/a}}
| 937
| 1,290
| 16
| 1,620
| 829.44
| 105,680
| 13,210
| 6,605
| 250
| PCIe FHHL card
|-
! style="text-align:left;" | T4 GPU accelerator (PCIe card)<ref name='T4'>{{cite news|title=NVIDIA TESLA T4 TENSOR CORE GPU|url=https://www.nvidia.com/en-us/data-center/tesla-t4/|access-date=17 October 2018|publisher=NVIDIA}}</ref><ref>{{cite web |url=https://www.nvidia.com/content/dam/en-zz/Solutions/Data-Center/tesla-t4/t4-tensor-core-product-brief.pdf |format=PDF |title=NVIDIA Tesla T4 Tensor Core Product Brief |website=www.nvidia.com |access-date=2019-07-10}}</ref>
| [[图灵微架构|Turing]]
| 2018年 9月12日
| 1× TU104-895-A1
| {{n/a}}
| 2,560
| 585
| 1,590
| GDDR6
| 256
| 16
| 5,000
| 320
| 64,800
| 8,100
| {{unk}}
| 7.5
| 70
| PCIe card
|-
! style="text-align:left;" | A2 GPU accelerator (PCIe card)<ref name='A2'>{{cite news|title=NVIDIA TESLA A2 TENSOR CORE GPU|url=https://www.nvidia.com/en-us/data-center/products/a2/ }}</ref>
| rowspan=6| [[安培微架构|Ampere]]
| 2021年 11月10日
| 1× GA107
| {{n/a}}
| 1,280
| 1,440
| 1,770
| GDDR6
| 128
| 16
| 6,252
| 200
| 18,124
| 4,531
| 140
| 8.6
| 40-60
| PCIe card (half height, single-slot)
|-
! style="text-align:left;" | A10 GPU accelerator (PCIe card)<ref name='A10'>{{cite news|title=NVIDIA TESLA A10 TENSOR CORE GPU|url=https://www.nvidia.com/en-us/data-center/products/a10-gpu/ }}</ref>
| 2021年 4月12日
| 1× GA102-890-A1
| {{n/a}}
| 9,216
| 885
| 1,695
| GDDR6
| 384
| 24
| 6,252
| 600
| 124,960
| 31,240
| 976
| 8.6
| 150
| PCIe card (single-slot)
|-
! style="text-align:left;" | A16 GPU accelerator (PCIe card)<ref name='A16'>{{cite news|title=NVIDIA TESLA A16 TENSOR CORE GPU|url=https://www.nvidia.com/en-us/data-center/products/a16-gpu/ }}</ref>
| 2021年 4月12日
| 4× GA107
| {{n/a}}
| 4× 1,280
| 885
| 1,695
| GDDR6
| 4× 128
| 4× 16
| 7,242
| 4× 200
| 4x 18,432
| 4× 4,608
| 1,084.8
| 8.6
| 250
| PCIe card (dual-slot)
|-
! style="text-align:left;" | A30 GPU accelerator (PCIe card)<ref name='A30'>{{cite news|title=NVIDIA TESLA A30 TENSOR CORE GPU|url=https://www.nvidia.com/en-us/data-center/products/a30-gpu/ }}</ref>
| 2021年 4月12日
| 1× GA100
| {{n/a}}
| 3,584
| 930
| 1,440
| HBM2
| 3,072
| 24
| 1,215
| 933.1
| 165,120
| 10,320
| 5,161
| 8.0
| 165
| PCIe card (dual-slot)
|-
! style="text-align:left;" | A40 GPU accelerator (PCIe card)<ref name='A40'>{{cite news|title=NVIDIA TESLA A40 TENSOR CORE GPU|url=https://www.nvidia.com/en-us/data-center/a40/ }}</ref>
| 2020年 10月5日
| 1× GA102
| {{n/a}}
| 10,752
| 1,305
| 1,740
| GDDR6
| 384
| 48
| 7,248
| 695.8
| 149,680
| 37,420
| 1,168
| 8.6
| 300
| PCIe card (dual-slot)
|-
! style="text-align:left;" | A100 GPU accelerator (PCIe card)<ref name='A100'>{{cite news|title=NVIDIA TESLA A100 TENSOR CORE GPU|url=https://www.nvidia.com/en-us/data-center/a100/ |access-date= 14 January 2021|publisher=NVIDIA}}</ref><ref>{{cite web |url=https://www.nvidia.com/content/dam/en-zz/Solutions/Data-Center/a100/pdf/A100-PCIE-Prduct-Brief.pdf |format=PDF |title=NVIDIA Tesla A100 Tensor Core Product Brief |website=www.nvidia.com |access-date=2020-09-22}}</ref>
| 2020年 5月14日<ref name="anand-A100">{{cite news|author=Smith|first=Ryan|date=May 14, 2020|title=NVIDIA Ampere Unleashed: NVIDIA Announces New GPU Architecture, A100 GPU, and Accelerator|publisher=AnandTech|url=https://www.anandtech.com/show/15801/nvidia-announces-ampere-architecture-and-a100-products}}</ref>
| 1× GA100-883AA-A1
| {{n/a}}
| 6,912
| 765
| 1410
| HBM2
| 5,120
| 40 or 80
| 1,215
| 1,555
| 312,000
| 19,500
| 9,700
| 8.0
| 250
| PCIe card (dual-slot)
|-
! style="text-align:left;" | H100 GPU accelerator (PCIe card)<ref>https://www.nvidia.com/en-us/data-center/h100/</ref>
| rowspan=2| [[Hopper (微架构)|Hopper]]
| rowspan=2| 2022年 3月22日<ref>https://wccftech.com/nvidia-hopper-gh100-gpu-official-5nm-process-worlds-fastest-hpc-chip-80-billion-transistors-hbm3-memory/</ref>
| rowspan=2| 1× GH100<ref>https://www.techpowerup.com/gpu-specs/h100-pcie.c3899</ref>
| {{n/a}}
| 14,592
| 1,065
| 1,755 CUDA 1620 TC
| HBM2e
| 5120
| 80
| 1,000
| 2,039
| 756,449
| 51,200
| 25,600
| 9.0
| 350
| PCIe card (dual-slot)
|-
! style="text-align:left;" | H100 GPU accelerator (SXM card)
| {{n/a}}
| 16,896
| 1,065
| 1,980 CUDA 1,830 TC
| HBM3
| 5,120
| 80
| 1,500
| 3,352
| 989,430
| 66,900
| 33,500
| 9.0
| 700
| [[SXM (socket)|SXM]] card
|-
! style="text-align:left;" | L40 GPU accelerator<ref>https://www.nvidia.com/en-us/data-center/l40/</ref>
| rowspan=2| [[Ada Lovelace (微架构)|Ada Lovelace]]
| 2022年 10月13日
| 1× AD102<ref>https://www.techpowerup.com/gpu-specs/l40.c3959</ref>
| {{n/a}}
| 18,176
| 735
| 2,490
| GDDR6
| 384
| 48
| 2,250
| 864
| 362,066
| 90,516
| 1,414
| 8.9
| 300
| PCIe card (dual-slot)
|-
! style="text-align:left;" | L4 GPU accelerator<ref>https://www.nvidia.com/en-us/data-center/l4/</ref><ref>https://images.nvidia.com/aem-dam/Solutions/Data-Center/l4/nvidia-ada-gpu-architecture-whitepaper-v2.1.pdf</ref>
| 2023年 3月21日<ref>https://investor.nvidia.com/news/press-release-details/2023/NVIDIA-and-Google-Cloud-Deliver-Powerful-New-Generative-AI-Platform-Built-on-the-New-L4-GPU-and-Vertex-AI/default.aspx</ref>
| 1x AD104<ref>https://www.techpowerup.com/gpu-specs/l4.c4091</ref>
| {{n/a}}
| 7,424
| 795
| 2,040
| GDDR6
| 192
| 24
| 1,563
| 300
| 121,000
| 30,300
| 490
| 8.9
| 72
| HHHL single slot PCIe card
|-
! rowspan=2 | 型号
! rowspan=2 | [[微架構]]
! rowspan=2 | 发布时间
! rowspan=2 | 芯片
! rowspan=2 | 核心 频率
! colspan=3 | 渲染器
! colspan=5 | 内存
! colspan=3 | 算力{{efn|name="Calculate"|group="NvidiaTesla}}
! rowspan=2 | CUDA 运算能力
! rowspan=2 | [[热设计功耗]]
! rowspan=2 | 注释/ 外形尺寸
|-
! CUDA核心 （总数）
! 基础 频率
! 最大 增强 频率{{efn|name="GPUBoost"|group="NvidiaTesla"}}
! 总线 类型
! 总线字长
! 容量大小
! 频率
! 带宽
! [[半精度浮点数|半精度]] Tensor Core FP32 累加
! [[單精度浮點數|單精度]] （[[乘積累加運算|MAD]]或[[积和熔加运算|FMA]]）
! [[雙精度浮點數|雙精度]] （[[积和熔加运算|FMA]]）
|}
*此外2023年3月21日以及2024年11月18日分別推出[[Hopper微架构]]的H800<ref name="q636">{{cite web | title=NVIDIA H800 SXM5 Specs | website=TechPowerUp | date=2025-04-18 | url=https://www.techpowerup.com/gpu-specs/h800-sxm5.c3975 | access-date=2025-04-18}}</ref>以及H200<ref name="q330">{{cite web | title=NVIDIA H200 NVL Specs | website=TechPowerUp | date=2025-04-18 | url=https://www.techpowerup.com/gpu-specs/h200-nvl.c4254 | access-date=2025-04-18}}</ref>
{{Sticky table end}}
{{navbar|Nvidia Tesla|mini=y}}
注释：
{{notelist|group=NvidiaTesla}}<noinclude>
{{Template reference list}}
[[Category:輝達圖形處理器]]
</noinclude>

Template:Nvidia Tesla - 版本历史

imported>Comrade John：​增加或調整內部連結

imported>Comrade John：增加或調整內部連結