Performance Rating
NVIDIA RTX PRO 6000 Blackwell Server outperforms NVIDIA A2 by 1,351.26% in the overall GPU ARK performance rating
NVIDIA RTX PRO 6000 Blackwell Server
Expert Comparison
NVIDIA A2 более экономичный вариант с меньшим количеством ядер и памяти, но достаточным для базовых задач, таких как визуализация и轻化版本： NVIDIA A2 是一个更经济的选择，适用于基本任务如可视化和图形处理，具有较少的核心数和内存。相比之下，NVIDIA RTX PRO 6000 Blackwell Server 配备更多核心和更大内存，适用于高性能计算和专业图形工作负载，性能远超 A2。
Contents:
Memory ML Performance Compute Power Architecture & Compatibility ML Software Support Clocks & Performance Power Consumption Rendering Benchmarks AdditionalMemory
Memory Size
+500%
96
ГБ
Memory Type
Memory Bandwidth
1.79 TB/s
Memory Bus Width
ML Performance
FP16 (Half Precision)
+2,681%
126.0
TFLOPS
BF16 (Brain Float)
TF32 (TensorFloat)
Compute Power
FP32 (Single Precision)
+2,681%
126.0
TFLOPS
FP64 (Double Precision)
+2,680%
1.968
TFLOPS
CUDA Cores
+1,780%
24,064
RT Cores
+1,780%
188
Architecture & Compatibility
GPU Architecture
SM (Streaming Multiprocessor)
+1,780%
188
PCIe Version
ML Software Support
CUDA Version
12.0
CUDA Toolkit (first supported)
CUDA Toolkit status
Clocks & Performance
Base Clock
+10%
1,590
Boost Clock
+48%
2,617
Memory Clock
+12%
1,750
Power Consumption
Recommended PSU
-75%
250 W
Power Connector
TDP/TGP
-90%
60 W
Rendering
Texture Units (TMU)
+1,780%
752
ROP
+1,780%
188
L2 Cache
2 MB
Benchmarks
MLPerf, llama2-70b-99.9 (fp4)
MLPerf, llama3.1-8b (fp4)
Geekbench AI, FP16
Geekbench AI, INT8
Geekbench AI, FP32
MLPerf, mixtral-8x7b (fp8)
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
