Performance Rating
NVIDIA GeForce RTX 5070 Ti SUPER outperforms NVIDIA Tesla T4 by 210.4% 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
16
ГБ
Memory Type
Memory Bandwidth
Memory Bus Width
ML Performance
FP16 (Half Precision)
43.94
TFLOPS
BF16 (Brain Float)
TF32 (TensorFloat)
Compute Power
FP32 (Single Precision)
+440%
43.94
TFLOPS
FP64 (Double Precision)
+170%
0.6866
TFLOPS
CUDA Cores
+250%
8,960
RT Cores
+75%
70
Architecture & Compatibility
GPU Architecture
SM (Streaming Multiprocessor)
+75%
70
PCIe Version
ML Software Support
CUDA Version
12.0
Clocks & Performance
Base Clock
+292%
2,295
Boost Clock
+54%
2,452
Memory Clock
+40%
1,750
Power Consumption
TDP/TGP
-80%
70 W
Recommended PSU
-67%
250 W
Power Connector
Rendering
Texture Units (TMU)
+75%
280
ROP
+75%
70
L2 Cache
+1,100%
48 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
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
