Live DB — analyst tables
Test
Amazon 32768 P4d-A100 NVL8 Nitro-v3 Hyperscaler EFA Cluster Configuration
AWS-P4d-A100 NVL8-EFA-WB
Amazon · A100 · Nitro-v3
Rack type
P4d NVL8
Cluster / location
P4d-A100 NVL8 Clusters
Buyer profile
Hyperscaler
Networking type
EFA
Number of XPUs
32,768
NIC used
Nitro-v3
NICs / XPU
0.5
Scale-out BW / XPU
50G
XPU / server
8
Scale-up network
1 rows
Assigned blockDefault
A100 - NVL8
Scale-UpXPU: A100 NVL81 lineshared by 9 SKUs
Scale-up Units are based on the scale-up world size (8 XPUs).
XPU World Size
2
WS| Scale-Up Network Layer | Part / Model | Units# | Unit BWGbit/s uni-d | Lane Count# | Lane BWGbit/s uni-d |
|---|---|---|---|---|---|
| XPUs per Node | · | 8 | 2,400 | 24 | 100 |
| Scale-Up World Size | · | 8 | 19,200 | · | · |
Switches
2
SW| Scale-Up Network Layer | Part / Model | Units# | Unit BWGbit/s uni-d | Lane Count# | Lane BWGbit/s uni-d |
|---|---|---|---|---|---|
| per Switch | Nvidia NVSwitch 2.0 (NVLink 3) | 6 | 2,400 | 24 | 100 |
| Scale-Up Switch Total | · | · | 14,400 | · | · |
Notes
- XPUs per Node — 600GB Bidi
Scale-Up Parts Summary
1
| Item | Qty/Cluster | Qty/Server | Attach/XPU |
|---|---|---|---|
| Nvidia NVSwitch 2.0 (NVLink 3) | 24,576 | 6.00 | 0.75 |
Backend network
5 rows
Assigned blockDefault
A100 - P4d NVL8 - EFAv2 - 3L - ETH - TH2 - AWS
BackendXPU: A100 P4d NVL85 linesshared by 1 SKU
Backend Units are based on the cluster size (32,768 XPUs).
Switches
5
SW Network Layer | Switch and Server Spec | Aggregate Port Utilization | Network Bandwidth Aggregate | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
SwitchPart Name / Model | Units# | Unit BWGbit/s | Port BWGbit/s | Downlink Ports / SW# | Uplink Ports / SW# | Max Ports# | Downlink Ports# | Uplink Ports# | Total Ports# | Downlink BWTbit/s | Uplink BWTbit/s | OversubRatio | |
| XPUs | · | 32,768 | 50 | 100 | 0 | 16,384 | 16,384 | 0 | 16,384 | 16,384 | 0 | 1,638.4 | · |
| NICs per Node | EFAv2 Nitro-v3 100G NIC | 4 | 100 | 100 | · | · | 8 | · | · | · | · | · | · |
| ToR | WhiteBox Tomahawk 2 6.4T | 512 | 6,400 | 100 | 32 | 32 | 64 | 16,384 | 16,384 | 32,768 | 1,638.4 | 1,638.4 | 1 |
| Leaf | WhiteBox Tomahawk 3 12.8T | 256 | 12,800 | 200 | 32 | 32 | 64 | 8,192 | 8,192 | 16,384 | 1,638.4 | 1,638.4 | 1 |
| Spine | WhiteBox Tomahawk 3 12.8T | 128 | 12,800 | 200 | 64 | 0 | 64 | 8,192 | 0 | 8,192 | 1,638.4 | 0 | · |
Transceivers
8
CN Network Layer | Transceiver / CablePart Name / Model | Total Units# | Units per SW/GPU# | Unit BWGbit/s | Lanes# | BW per LaneGbit/s | Agg. BWTbit/s |
|---|---|---|---|---|---|---|---|
| GPU Nodes | None - AECs are Used -> | 32,768 | 32,768 | 50 | 1 | 50 | 1,638.4 |
| CPU Nodes | None - AECs are Used -> | 8,192 | · | · | · | · | · |
| Leaf | |||||||
| Downlink | Generic 200G DR4 QSFP56 | 8,192 | 32 | 200 | 4 | 50 | 1,638.4 |
| Uplink | Generic 200G DR4 QSFP56 | 8,192 | #ERR | 200 | #ERR | 50 | 0 |
| Spine | |||||||
| Downlink | Generic 200G DR4 QSFP56 | 8,192 | #ERR | 200 | #ERR | 50 | 0 |
| Uplink | · | · | · | · | · | · | · |
| Core | |||||||
| Downlink | · | · | · | · | · | · | · |
| Uplink | · | · | · | · | · | · | · |
| Super-Core | |||||||
| ToR | |||||||
| Downlink | None - AECs are Used -> | 16,384 | 16 | 100 | 4 | 25 | 1,638.4 |
| Uplink | Generic 200G DR4 QSFP56 | 16,384 | 16 | 100 | 4 | 25 | 1,638.4 |
Cables
8
CB Network Layer | Cable/ FiberType | Total Units# | Units per Transceiver# | Unit BWGbit/s | Lane Count# | SerDes BWGbit/s |
|---|---|---|---|---|---|---|
| GPU Nodes | Generic 200G Y Cable AEC | 16,384 | 0.5 | 100 | 2 | 50 |
| CPU Nodes | · | · | · | · | · | · |
| Leaf | ||||||
| Downlink | Already counted above | · | · | · | · | · |
| Uplink | Generic 200G Single Mode | 8,192 | 1 | 200 | 4 | 50 |
| Spine | ||||||
| Downlink | · | · | · | · | · | · |
| Uplink | · | · | · | · | · | · |
| Core | ||||||
| Downlink | · | · | · | · | · | · |
| Uplink | · | · | · | · | · | · |
| Super-Core | ||||||
| ToR | ||||||
| Downlink | Already counted above | · | · | · | · | · |
| Uplink | Generic 200G Single Mode | 8,192 | 0.5 | 200 | 4 | 50 |
Backend Parts Summary
5
| Item | Qty/Cluster | Qty/Server | Attach/XPU |
|---|---|---|---|
| Generic 200G Y Cable AEC | 16,384 | 4.00 | 0.50 |
| Generic 200G DR4 QSFP56 | 40,960 | 10.00 | 1.25 |
| Generic 200G Single Mode | 16,384 | 4.00 | 0.50 |
| WhiteBox Tomahawk 2 6.4T | 512 | 0.13 | 0.02 |
| WhiteBox Tomahawk 3 12.8T | 384 | 0.09 | 0.01 |
Frontend network
No converged frontend + storage fabric data for this cluster.
Frontend Parts Summary
0
| Item | Qty/Cluster | Qty/Server | Attach/XPU |
|---|
Compute Tray Networking Content
2 rows
Networking components inside the compute tray — accelerators, board-to-board connectors, retimers, and their per-tray bandwidth.
Assigned blockDefault
A100 - P4d NVL8 - AWS
Compute TrayXPU: A100 P4d NVL82 linesshared by 1 SKU
| Tray layer | Part / Model | Units per Tray | Unit BWGb/s | Lanes | Lane BWGb/s | Per switch |
|---|---|---|---|---|---|---|
| XPUs per Tray | Blackwell 200 | 8 | · | · | · | · |
| Compute Tray | Broadcom PEX89144 Gen 5.0 144L | 4 | · | · | · | · |
| GPU to CPU Tray | Generic PCIe Gen5 x16 2-meter DAC | 8 | · | · | · | · |
Compute Tray Parts Summary
2
| Item | Qty/Cluster | Qty/Server | Attach/XPU |
|---|---|---|---|
| Broadcom PEX89144 Gen 5.0 144L | 16,384 | 4.00 | 0.50 |
| Generic PCIe Gen5 x16 2-meter DAC | 32,768 | 8.00 | 1.00 |