COMPUTE ATLAS Supercomputer supply-chain graph
1581 systems 558 sites Sourced data

Supercomputers/Systems/APEX (CMKL)

Thailand · operational · ai training

APEX (CMKL)

RSS for this system Your watchlist

Also known as CMKL DGX POD; CMKL University AI cluster

Operated by CMKL University .

Researched with AI assistance, checked by scripts. These rows were researched by AI agents reading public operator, laboratory and vendor sources, appended by scripts and checked by automated validation; no subject-matter expert has reviewed them (how this site is made). A mechanical second-reader pass has since fetched every cited source: 1350 of 1581 systems have a readable citation that names them, and 64 are genuinely weakly sourced. Every claim carries its source and a confidence tier. Treat anything below verified as a lead, not a citation.

Measured performance

Rmax
-
Rpeak
-
Rmax ÷ Rpeak
-
Cores
-
Power
-
Per watt
- GF/W

Figures are as last publicly reported for the configuration described below, not a live measurement. Where a system was upgraded in place, the post-upgrade configuration is shown and the earlier one appears in the timeline.

What this machine is made of

One row per supplier relationship. “Supplier at build” is the company that shipped the part at the time; where that company has since been acquired, the parent it rolls up to today is shown beside it. That distinction is what makes ticker-level aggregation possible across a thirty-year dataset.

Role Part Supplier at build Quantity Confidence Source
Integrator -
TCC Technology integrated data centre, network, GPU and storage infrastructure for APEX at CMKL University
 TCC Technology
- Reported tcc-technology.com
Accelerator NVIDIA A100 (in DGX A100)
Accelerators
Six NVIDIA DGX A100 systems (8 A100 GPUs each per the DGX A100 configuration)
 NVIDIA
NVDANASDAQ
48
estimated
 Reported nvidia.comddn.com
Interconnect -
NVIDIA Mellanox Spectrum 100GbE switches with Cumulus software and Quantum 200 Gb/s InfiniBand switches
 NVIDIA
NVDANASDAQ
- Reported nvidia.com
Storage -
Two DDN AI400X all-NVMe appliances (A3I), 645 TB raw flash plus 3.2 PB raw disk per DDN
 DDN
2
reported
 Reported ddn.comnvidia.com

The read

A DGX POD style cluster at CMKL University, the Carnegie Mellon and KMITL collaboration in Thailand, announced by NVIDIA on 13 October 2020: six NVIDIA DGX A100 systems quoted at 30 petaflops of AI performance, linked by NVIDIA Cumulus software, Spectrum 100GbE switches and Quantum 200 Gb/s InfiniBand. DDN's own release says the storage is two all-NVMe DDN AI400X appliances with 645 TB raw flash and 3.2 PB raw disk, while NVIDIA's page says 2.5 PB of DDN storage; the two figures are not reconciled here. NVIDIA expected the pod to be operational in early November 2020, and a 1 April 2021 release from integrator TCC Technology names the cluster APEX, located at CMKL, as national AI research infrastructure for Thailand's National Data Platform for AI, backed by the Ministry of Higher Education, Science, Research and Innovation and PMU-C (NVIDIA cites NXPO and PMU-C funding). Planned workloads were PyTorch image processing, Kaldi speech recognition and Metropolis DeepStream video analytics. The 48 A100 GPUs in the parts list are six systems times the eight GPUs of a DGX A100, because the CMKL pages give only the system count. The start-of-service date is not confirmed and no source after July 2021 was found. The larger Thai national machine is lanta.

Timeline

Change history

Same accelerator

Further reading

Operating a supercomputer

Acceptance testing, why "installed" and "in production" are six months apart, and the machine lifecycle.

Inside a node

Why accelerators exist, what the memory hierarchy costs, and how the CPU and accelerator merged onto one package.

Storage and I/O

Parallel file systems, why checkpointing dominates the write load, and the metadata failure mode nobody expects.