blackm3285
blackm3285
  • Видео 286
  • Просмотров 2 660
TYAN TN70A-B8026 - Budget EPYC 7001/7002 system - with S8026 motherboard
VRM temperture reach over 75 °C in 3 min R23 test with mini heatsink and 200W cpu, room temperature at 26°C
Here is the specific part list of this machine
Model:TYAN Transport SX TN70A-B8026
CPU: AMD EPYC 7532
Cooler: TYAN 343T57100001 180W 1U Passive
RAM: ATP X4B16QC4BNVFME-E-FO1 1Rx4 PC4-2933Y 16GB RDIMM *16
MB: TYAN Tomcat SX S8026GM2NRE
GPU: ASPEED AST2500 Graphics DDR4
   Huawei 06320105 NVIDIA Tesla P100 PCIe 16GB HBM2 ECC
   HPE 868199-001 NVIDIA Tesla P100 PCIe 16GB HBM2 ECC
SSD (M.2): Samsung PM983 960GB -- RAID0
     SK Hynix PE4010 960GB -- RAID0
     DELL 0J9GY6 LITEON EPX-KW960 960GB -- RAID0
     HPE 869255-002 Seagate Nytro XM1440 800GB
SSD (U.2): Intel SSD DC P3700 400GB
     Lenovo SSS7...
Просмотров: 75

Видео

Quad NVIDIA Tesla P40 - Stable Diffusion Benchmark - Budget 24GB GPU for AI
Просмотров 11514 дней назад
Following tests are with SwarmUI Frontend and ComfyUI Backend : 1. SD 1.5, 512x768, 25 steps, DPM 2M Karras 2. SD 1.5, 512x768 upscale to 1024x1536, Denoising strength 0.5 (13 steps), DPM 2M Karras 3. SDXL (Pony), 1024x1536, 20 steps, DPM 2M Karras Models are stored in Samsung PM1725b 1.6TB AIC Result : 1. Speed : 2.19it/s , Time : 11s , -9.6% slower than P100, Peak VRAM 5487MB 2. Speed : 3.47s...
Intel Optane DCPMM (AEP) 2* 512GB vs DC P4800X 375GB - Photoshop Speed Testing (銀河圖展開)
Просмотров 4421 день назад
Intel Optane DCPMM (AEP) 2* 512GB vs DC P4800X 375GB - Photoshop Speed Testing (銀河圖展開)
Intel Optane SSD DC P4800X 375GB U.2 - CrystalDiskMark Speed Testing - High-End Enterprice SSD
Просмотров 3921 день назад
Intel Optane SSD DC P4800X 375GB U.2 - CrystalDiskMark Speed Testing - High-End Enterprice SSD
Dual NVIDIA Tesla P100 PCIe 16GB - Stable Diffusion Benchmark - Budget 16GB HBM GPU for AI
Просмотров 182Месяц назад
Dual NVIDIA Tesla P100 PCIe 16GB - Stable Diffusion Benchmark - Budget 16GB HBM GPU for AI
Quad NVIDIA Tesla P40 - Cinebench 2024 GPU Run - Budget 24GB GPU for AI
Просмотров 104Месяц назад
Quad NVIDIA Tesla P40 - Cinebench 2024 GPU Run - Budget 24GB GPU for AI
Dual Intel Xeon E5-2696 v4 - Cinebench R23 Multi-Core Run - Best LGA2011-3 OEM CPU
Просмотров 95Месяц назад
Dual Intel Xeon E5-2696 v4 - Cinebench R23 Multi-Core Run - Best LGA2011-3 OEM CPU
Dual NVIDIA Tesla P100 PCIe 16GB - Cinebench 2024 GPU Run - Budget 16GB HBM GPU for AI
Просмотров 116Месяц назад
Dual NVIDIA Tesla P100 PCIe 16GB - Cinebench 2024 GPU Run - Budget 16GB HBM GPU for AI
Dual AMD Radeon Pro W5700 - Cinebench 2024 GPU Run - Budget Workstation GPU
Просмотров 64Месяц назад
Dual AMD Radeon Pro W5700 - Cinebench 2024 GPU Run - Budget Workstation GPU
AMD EPYC 7532 - Cinebench R23 Multi-Core Run - Budget 32 Core EPYC CPU
Просмотров 156Месяц назад
AMD EPYC 7532 - Cinebench R23 Multi-Core Run - Budget 32 Core EPYC CPU
Intel Optane DCPMM (AEP) 2* 512GB AD Mode - CrystalDiskMark Speed Testing - High capacity DDR4 DIMMs
Просмотров 103Месяц назад
Intel Optane DCPMM (AEP) 2* 512GB AD Mode - CrystalDiskMark Speed Testing - High capacity DDR4 DIMMs
Intel Xeon Platinum 8275CL - Cinebench R23 Multi-Core Run - Best LGA3647 OEM CPU
Просмотров 184Месяц назад
Intel Xeon Platinum 8275CL - Cinebench R23 Multi-Core Run - Best LGA3647 OEM CPU

Комментарии

  • @repixelatedmc
    @repixelatedmc 6 дней назад

    Quick question, why even a p40 if it needs to run on fp32 and has equivalent performance of a 3060 that is able to run on fp16 ( 2x less vram )

    • @blackm3285
      @blackm3285 6 дней назад

      ​​@@repixelatedmc because its cheaper, I bought these card as 820 CNY (115 USD) per card, but now it become double of this price. Also, 24GB of physical VRAM let it more future-proof.

    • @blackm3285
      @blackm3285 6 дней назад

      By the way, using FP16 on GPU can't use less VRAM, only become faster, because the VRAM requirements is depending on models you using, and the SD models you can find now usually are "purned FP16", if your GPU don't support FP16, it can be load on GPU and run it as FP32 by stuffing zero on missing precision. I tested this when community still provide FP32 and FP16 models, and using FP32 model will become a little bit faster.

  • @CornBreadMan264
    @CornBreadMan264 18 дней назад

    What's the AVX-512 All-core ratio? I noticed the non-AVX all core is 3.6GHz and the AVX2 is 3.4GHz

    • @blackm3285
      @blackm3285 18 дней назад

      @@CornBreadMan264 In R23 is AVX2 light, rated at 3.6GHz, but it throttled to 3.4GHz, AVX2 heavy is rated at 3.1GHz, you can see this frequency in R24, AVX512 is at 2.7GHz

  • @sigitpambudi3923
    @sigitpambudi3923 28 дней назад

    So could we use dcpmm as regular ram to run apps?

    • @blackm3285
      @blackm3285 28 дней назад

      Yes we can, just get into BIOS, then set as Memory Mode, it will use DCPMM as DRAM, use DDR4 DIMMs as something like L4 cache

    • @sigitpambudi3923
      @sigitpambudi3923 28 дней назад

      @@blackm3285 what will happened when we use DDR4 as l4 cache

    • @blackm3285
      @blackm3285 28 дней назад

      ​@@sigitpambudi3923 In my system will show total 1.2TB of DDR4, 1004GB can use, 192GB reserved, and you can use these 1TB as normal, but when you use more than 192GB, you may feel laggy sometimes

    • @sigitpambudi3923
      @sigitpambudi3923 28 дней назад

      @@blackm3285 lets say someone use 1:4 , 1,5tb ram and 6tb dcpmm, so that 6tb dcpmm could works like ram?

    • @blackm3285
      @blackm3285 28 дней назад

      @@sigitpambudi3923 Yes

  • @DeltaTacticalGamer
    @DeltaTacticalGamer Месяц назад

    what is this or is a AI Benchmark

    • @blackm3285
      @blackm3285 Месяц назад

      Its an AI image generation software called "Stable Diffusion", I tested three different type of common work load for this

  • @luchi3483
    @luchi3483 Месяц назад

    Nice! You can push some extra points with software optimization and using real time processing priority. I just got a nice pair of 2698 v4 for 115$ on ebay hopefully I will get around 30000 points on cinebench

  • @kimjang_
    @kimjang_ Месяц назад

    Wow, it's really the second generation of the last king oem cpu. Doesn't it support 2cpu?

    • @blackm3285
      @blackm3285 Месяц назад

      it only support "up to" 2 cpus, most of platinum cpu can support up to 8 cpus, i only test one because this motherboard only has one socket.