Skip to content
Snippets Groups Projects
Commit be3f40ce authored by Jan Siwiec's avatar Jan Siwiec
Browse files

OJ proofread

parent 1606966f
Branches
Tags
No related merge requests found
Pipeline #30506 passed with warnings
Stage: test
Stage: build
Stage: deploy
Stage: after_test
Hide whitespace changes
Inline Side-by-side
docs.it4i/barbora/compute-nodes.md
+ 34
− 34
View file @ be3f40ce
...@@ -6,13 +6,13 @@ Barbora is a cluster of x86-64 Intel-based nodes built with the BullSequana Comp ...@@ -6,13 +6,13 @@ Barbora is a cluster of x86-64 Intel-based nodes built with the BullSequana Comp
* 192 nodes * 192 nodes
* 6912 cores in total * 6912 cores in total
* 2x Intel Cascade Lake 6240, 18-core, 2.6 GHz processors per node * 2x Intel Cascade Lake 6240, 18-core, 2.6GHz processors per node
* 192 GB DDR4 2933MT/s of physical memory per node (12x 16 GB) * 192GB DDR4 2933MT/s of physical memory per node (12x16GB)
* BullSequana X1120 blade servers * BullSequana X1120 blade servers
* 2995,2 GFLOP/s per compute node * 2995.2 GFLOP/s per compute node
* 1x 1 GB Ethernet * 1x 1GB Ethernet
* 1x HDR100 IB port * 1x HDR100 IB port
* 3 computes nodes per X1120 blade server * 3 compute nodes per X1120 blade server
* cn[1-192] * cn[1-192]
![](img/BullSequanaX1120.png) ![](img/BullSequanaX1120.png)
...@@ -21,13 +21,13 @@ Barbora is a cluster of x86-64 Intel-based nodes built with the BullSequana Comp ...@@ -21,13 +21,13 @@ Barbora is a cluster of x86-64 Intel-based nodes built with the BullSequana Comp
* 8 nodes * 8 nodes
* 192 cores in total * 192 cores in total
* two Intel Skylake Gold 6126, 12-core, 2.6 GHz processors per node * two Intel Skylake Gold 6126, 12-core, 2.6GHz processors per node
* 192 GB DDR4 2933MT/s with ECC of physical memory per node (12x 16 GB) * 192 GB DDR4 2933MT/s with ECC of physical memory per node (12x16GB)
* 4x GPU accelerator NVIDIA Tesla V100-SXM2 per node * 4x GPU accelerator NVIDIA Tesla V100-SXM2 per node
* Bullsequana X410-E5 NVLink-V blade servers * Bullsequana X410-E5 NVLink-V blade servers
* 1996,8 GFLOP/s per compute nodes * 1996.8 GFLOP/s per compute nodes
* GPU-tp-GPU All-to-All NVLINK 2.0, GPU-Direct * GPU-to-GPU All-to-All NVLINK 2.0, GPU-Direct
* 1 GB Ethernet * 1GB Ethernet
* 2x HDR100 IB ports * 2x HDR100 IB ports
* cn[193-200] * cn[193-200]
...@@ -37,8 +37,8 @@ Barbora is a cluster of x86-64 Intel-based nodes built with the BullSequana Comp ...@@ -37,8 +37,8 @@ Barbora is a cluster of x86-64 Intel-based nodes built with the BullSequana Comp
* 1x BullSequana X808 server * 1x BullSequana X808 server
* 128 cores in total * 128 cores in total
* 8 Intel Skylake 8153, 16-core, 2.0 GHz, 125W * 8 Intel Skylake 8153, 16-core, 2.0GHz, 125W
* 6144 GiB DDR4 2667MT/s of physical memory per node (92x 64 GB) * 6144 GiB DDR4 2667MT/s of physical memory per node (92x64GB)
* 2x HDR100 IB port * 2x HDR100 IB port
* 8192 GFLOP/s * 8192 GFLOP/s
* cn[201] * cn[201]
...@@ -47,11 +47,11 @@ Barbora is a cluster of x86-64 Intel-based nodes built with the BullSequana Comp ...@@ -47,11 +47,11 @@ Barbora is a cluster of x86-64 Intel-based nodes built with the BullSequana Comp
## Compute Node Summary ## Compute Node Summary
| Node type | Count | Range | Memory | Cores | Queues | | Node type | Count | Range | Memory | Cores |
| ---------------------------- | ----- | ----------- | ------ | ----------- | -------------------------- | | ---------------------------- | ----- | ----------- | ------ | ----------- |
| Nodes without an accelerator | 192 | cn[1-192] | 192GB | 36 @ 2.6 GHz | qexp, qprod, qlong, qfree | | Nodes without an accelerator | 192 | cn[1-192] | 192GB | 36 @ 2.6GHz |
| Nodes with a GPU accelerator | 8 | cn[193-200] | 192GB | 24 @ 2.6 GHz | qnvidia | | Nodes with a GPU accelerator | 8 | cn[193-200] | 192GB | 24 @ 2.6GHz |
| Fat compute nodes | 1 | cn[201] | 6144GiB | 128 @ 2.0 GHz | qfat | | Fat compute nodes | 1 | cn[201] | 6144GiB | 128 @ 2.0GHz |
## Processor Architecture ## Processor Architecture
...@@ -116,23 +116,23 @@ Barbora is equipped with an [NVIDIA Tesla V100-SXM2][g] accelerator. ...@@ -116,23 +116,23 @@ Barbora is equipped with an [NVIDIA Tesla V100-SXM2][g] accelerator.
![](img/gpu-v100.png) ![](img/gpu-v100.png)
|NVIDIA Tesla V100-SXM2|| | NVIDIA Tesla V100-SXM2 | |
| --- | --- | | ---------------------------- | -------------------------------------- |
| GPU Architecture | NVIDIA Volta | | GPU Architecture | NVIDIA Volta |
| NVIDIA Tensor| Cores: 640 | | NVIDIA Tensor Cores | 640 |
| NVIDIA CUDA® Cores | 5 120 | | NVIDIA CUDA® Cores | 5120 |
| Double-Precision Performance | 7.8 TFLOP/s | | Double-Precision Performance | 7.8TFLOP/s |
| Single-Precision Performance | 15.7 TFLOP/s | | Single-Precision Performance | 15.7TFLOP/s |
| Tensor Performance | 125 TFLOP/s | | Tensor Performance | 125TFLOP/s |
| GPU Memory | 16 GB HBM2 | | GPU Memory | 16GB HBM2 |
| Memory Bandwidth | 900 GB/sec | | Memory Bandwidth | 900GB/sec |
| ECC | Yes | | ECC | Yes |
| Interconnect Bandwidth | 300 GB/sec | | Interconnect Bandwidth | 300GB/sec |
| System Interface | NVIDIA NVLink | | System Interface | NVIDIA NVLink |
| Form Factor | SXM2 | | Form Factor | SXM2 |
| Max Power Consumption | 300 W | | Max Power Consumption | 300W |
| Thermal Solution | Passive | | Thermal Solution | Passive |
| Compute APIs | CUDA, DirectCompute,OpenCLTM, OpenACC | | Compute APIs | CUDA, DirectCompute, OpenCLTM, OpenACC |
[a]: https://en.wikichip.org/wiki/intel/microarchitectures/skylake_(server)#Core [a]: https://en.wikichip.org/wiki/intel/microarchitectures/skylake_(server)#Core
[b]: https://en.wikichip.org/wiki/intel/microarchitectures/skylake_(client)#Pipeline [b]: https://en.wikichip.org/wiki/intel/microarchitectures/skylake_(client)#Pipeline
......
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Please register or to comment