NVIDIA H100 SXM5
About GPU
The NVIDIA H100 SXM5 GPU is a powerhouse of a graphics processing unit, specifically designed for professional use. With a base clock speed of 1590MHz and a boost clock speed of 1980MHz, this GPU offers impressive performance for demanding workloads. The 80GB of HBM3 memory and a memory clock of 1313MHz ensure that the H100 SXM5 can handle large datasets and complex calculations with ease.
One of the most impressive features of the H100 SXM5 is its 16896 shading units, allowing for highly parallel processing and efficient handling of complex graphics and compute workloads. Additionally, with 50MB of L2 cache, this GPU is capable of handling large amounts of data with low latency, further enhancing its overall performance.
In terms of power consumption, the H100 SXM5 has a TDP of 700W, which is on the higher end, but it is to be expected considering its high performance capabilities. With a theoretical performance of 66.91 TFLOPS, this GPU is well-suited for professional applications such as AI, deep learning, scientific simulations, and high-performance computing.
Overall, the NVIDIA H100 SXM5 GPU is a top-of-the-line solution for professionals and organizations in need of cutting-edge graphics and compute performance. Its impressive specifications and performance capabilities make it an excellent choice for those who require the utmost in processing power for their workloads.
Basic
Label Name
NVIDIA
Platform
Professional
Launch Date
March 2022
Model Name
H100 SXM5
Generation
Tesla Hopper
Base Clock
1590MHz
Boost Clock
1980MHz
Bus Interface
PCIe 5.0 x16
Transistors
80,000 million
Tensor Cores
?
Tensor Cores are specialized processing units designed specifically for deep learning, providing higher training and inference performance compared to FP32 training. They enable rapid computations in areas such as computer vision, natural language processing, speech recognition, text-to-speech conversion, and personalized recommendations. The two most notable applications of Tensor Cores are DLSS (Deep Learning Super Sampling) and AI Denoiser for noise reduction.
528
TMUs
?
Texture Mapping Units (TMUs) serve as components of the GPU, which are capable of rotating, scaling, and distorting binary images, and then placing them as textures onto any plane of a given 3D model. This process is called texture mapping.
528
Foundry
TSMC
Process Size
4 nm
Architecture
Hopper
Memory Specifications
Memory Size
80GB
Memory Type
HBM3
Memory Bus
?
The memory bus width refers to the number of bits of data that the video memory can transfer within a single clock cycle. The larger the bus width, the greater the amount of data that can be transmitted instantaneously, making it one of the crucial parameters of video memory. The memory bandwidth is calculated as: Memory Bandwidth = Memory Frequency x Memory Bus Width / 8. Therefore, when the memory frequencies are similar, the memory bus width will determine the size of the memory bandwidth.
5120bit
Memory Clock
1313MHz
Bandwidth
?
Memory bandwidth refers to the data transfer rate between the graphics chip and the video memory. It is measured in bytes per second, and the formula to calculate it is: memory bandwidth = working frequency × memory bus width / 8 bits.
3350 GB/s
Theoretical Performance
Pixel Rate
?
Pixel fill rate refers to the number of pixels a graphics processing unit (GPU) can render per second, measured in MPixels/s (million pixels per second) or GPixels/s (billion pixels per second). It is the most commonly used metric to evaluate the pixel processing performance of a graphics card.
47.52 GPixel/s
Texture Rate
?
Texture fill rate refers to the number of texture map elements (texels) that a GPU can map to pixels in a single second.
1045 GTexel/s
FP16 (half)
?
An important metric for measuring GPU performance is floating-point computing capability. Half-precision floating-point numbers (16-bit) are used for applications like machine learning, where lower precision is acceptable. Single-precision floating-point numbers (32-bit) are used for common multimedia and graphics processing tasks, while double-precision floating-point numbers (64-bit) are required for scientific computing that demands a wide numeric range and high accuracy.
267.6 TFLOPS
FP64 (double)
?
An important metric for measuring GPU performance is floating-point computing capability. Double-precision floating-point numbers (64-bit) are required for scientific computing that demands a wide numeric range and high accuracy, while single-precision floating-point numbers (32-bit) are used for common multimedia and graphics processing tasks. Half-precision floating-point numbers (16-bit) are used for applications like machine learning, where lower precision is acceptable.
33.45 TFLOPS
FP32 (float)
?
An important metric for measuring GPU performance is floating-point computing capability. Single-precision floating-point numbers (32-bit) are used for common multimedia and graphics processing tasks, while double-precision floating-point numbers (64-bit) are required for scientific computing that demands a wide numeric range and high accuracy. Half-precision floating-point numbers (16-bit) are used for applications like machine learning, where lower precision is acceptable.
68.248
TFLOPS
Miscellaneous
SM Count
?
Multiple Streaming Processors (SPs), along with other resources, form a Streaming Multiprocessor (SM), which is also referred to as a GPU's major core. These additional resources include components such as warp schedulers, registers, and shared memory. The SM can be considered the heart of the GPU, similar to a CPU core, with registers and shared memory being scarce resources within the SM.
132
Shading Units
?
The most fundamental processing unit is the Streaming Processor (SP), where specific instructions and tasks are executed. GPUs perform parallel computing, which means multiple SPs work simultaneously to process tasks.
16896
L1 Cache
256 KB (per SM)
L2 Cache
50MB
TDP
700W
Vulkan Version
?
Vulkan is a cross-platform graphics and compute API by Khronos Group, offering high performance and low CPU overhead. It lets developers control the GPU directly, reduces rendering overhead, and supports multi-threading and multi-core processors.
N/A
OpenCL Version
3.0
OpenGL
N/A
DirectX
N/A
CUDA
9.0
Power Connectors
8-pin EPS
Shader Model
N/A
ROPs
?
The Raster Operations Pipeline (ROPs) is primarily responsible for handling lighting and reflection calculations in games, as well as managing effects like anti-aliasing (AA), high resolution, smoke, and fire. The more demanding the anti-aliasing and lighting effects in a game, the higher the performance requirements for the ROPs; otherwise, it may result in a sharp drop in frame rate.
24
Suggested PSU
1100W
Benchmarks
FP32 (float)
Score
68.248
TFLOPS
Compared to Other GPU
FP32 (float)
/ TFLOPS