Performance Rating
NVIDIA B300 outperforms NVIDIA Tesla T4 by 1,016.93% in the overall GPU ARK performance rating
NVIDIA Tesla T4
Contents:
Memory ML Performance Compute Power Architecture & Compatibility ML Software Support Clocks & Performance Power Consumption Rendering Benchmarks AdditionalMemory
Memory Size
+800%
144
ГБ ×2 (288 ГБ)
Memory Type
Memory Bandwidth
4.10 TB/s
×2 (8.2 TB/s)
Memory Bus Width
ML Performance
FP16 (Half Precision)
1.2318
TFLOPS
BF16 (Brain Float)
TF32 (TensorFloat)
Compute Power
FP32 (Single Precision)
+846%
76.99
TFLOPS
FP64 (Double Precision)
+373%
1.2029
TFLOPS
CUDA Cores
+640%
18,944
×2 (37888)
RT Cores
Architecture & Compatibility
GPU Architecture
SM (Streaming Multiprocessor)
+270%
148
PCIe Version
ML Software Support
CUDA Version
10.3
Clocks & Performance
Base Clock
+185%
1,665
Boost Clock
+28%
2,032
Memory Clock
+60%
2,000
Power Consumption
Recommended PSU
-86%
250 W
Power Connector
TDP/TGP
-95%
70 W
Rendering
Texture Units (TMU)
+270%
592
×2 (1184)
ROP
L2 Cache
+1,150%
50 MB
Benchmarks
LLM Perf, Llama-2-7b-hf (bnb)
LLM Perf, Llama-2-7b-hf (gptq)
LLM Perf, Llama-2-7b-hf (unquantized)
LLM Perf, Llama-3.1-8B-Instruct (awq)
LLM Perf, Llama-3.1-8B-Instruct (bnb)
LLM Perf, Llama-3.1-8B-Instruct (gptq)
LLM Perf, Meta-Llama-3-8B (awq)
LLM Perf, Meta-Llama-3-8B (bnb)
LLM Perf, Meta-Llama-3-8B (gptq)
LLM Perf, Meta-Llama-3-8B-Instruct (awq)
LLM Perf, Meta-Llama-3-8B-Instruct (bnb)
LLM Perf, Meta-Llama-3-8B-Instruct (gptq)
llama.cpp, llama-2-7b-Q4_0
LLM Perf, Mistral-7B-Instruct-v0.2 (gptq)
LLM Perf, Mistral-7B-v0.1 (bnb)
LLM Perf, Mistral-7B-v0.1 (gptq)
LLM Perf, Mistral-7B-v0.1 (unquantized)
Additional
Slots
SXM Module
Release Date
Display Outputs
Renting is cheaper than buying
H100 80Gb SXM (80GB)
NVIDIA
Флагманская GPU на архитектуре Hopper с 80 ГБ памяти HBM3 в форм-факторе SXM. Создана для обучения и инференса больших языковых моделей, HPC задач. Поддерживает FP64, FP32, TF32, FP16, BF16, FP8 и INT8 вычисления с Tensor Cores 4-го поколения.
from 299.11 RUB/hour
