We take cryptocurrency — so the coins you hold can buy real hardware.Pay in crypto, spend it on real hardware. Free insured shipping worldwide over $399. Plain, unbranded boxes, sent from Germany.Free insured shipping over $399. Your rate is locked for 30 minutes once checkout opens.Rate locked 30 minutes. Every item sealed, serial-checked and covered by a 24-month warranty.Sealed, 24-month warranty.

Saved items Sign in
Eight green DDR5 RDIMM modules in anti-static tray

NEMIX RAM

NEMIX RAM 512GB (8x64GB) DDR5 4800 CL40 2Rx4 ECC RDIMM for THINKMATE GPX-XH8-22S4-8GPU

SKU CH-C179-JWSV3

Eight 64 GB DDR5 4800 CL40 2Rx4 ECC RDIMM modules totalling 512 GB at 1.1 V for the THINKMATE GPX-XH8-22S4-8GPU server

  • Total capacity512 GB (8 × 64 GB)
  • Rated speedDDR5 4800 (PC5 38400)
  • ChipsetMajor OEM Samsung, Micron, Hynix
  • CAS latencyCL40
  • Buffer typeRegistered
  • Rated voltage1.1 V
  • 512GB capacity scales GPU-dense workloads
  • 8 matched 64GB modules enable full-channel bandwidth
  • DDR5 4800 speed feeds eight GPUs with high throughput
  • ECC registered design protects data integrity under load
  • 10V operation reduces power draw at scale

Targeted Server Memory Kit

This NEMIX RAM kit delivers 512GB of DDR5 ECC registered memory across eight 64GB modules. It is built for the THINKMATE GPX-XH8-22S4-8GPU GPU-optimised server and suits IT teams running memory-intensive workloads that require error correction and registered signalling.

Build Integration

Each module is a 288-pin DDR5 RDIMM operating at 1.1V with a CL40 latency and 2Rx4 organisation. The registered design reduces electrical load on the server memory controller, freeing the platform to populate all slots without signal degradation. No heat spreaders are fitted, so airflow in the chassis must accommodate bare modules.

Upgrade Context

The kit replaces lower-capacity or non-ECC DDR5 RDIMMs when the workload demands 512GB of registered memory with on-die error correction. The jump is worth it when the application scales beyond the limits of smaller kits and the platform supports the full 8-module population at 4800 MT/s.

How it compares

Measured against everything we stock in Model-Specific Memory Upgrades, at today’s prices. Recalculated every night.

Price $13,268.99
Pricier than about 90% of the 21,161 comparable items (median $397.99).
Total capacity 512 GB (8 × 64 GB)
Higher than about 90% of the 20,941 comparable items (median 32 GB).
Price per GB $25.92 per GB
Pricier than about 71% of the 20,941 comparable items (median $8.19 per GB).

New to this? Buying memory and SSDs with crypto — what to check, what it costs to pay, and what is never asked for.

Specifications

Product
Brand nameNEMIX RAM
Product lineTHINKMATE Server/Workstation Compatible Series Memory
Model nameTHINKMATE DDR5 4800 MHz ECC RDIMM Registered Server Memory
Part numberMR38400-924K8
Total capacity512 GB (8 × 64 GB)
Product type288-Pin DDR5 RDIMM
Performance
Rated speedDDR5 4800 (PC5 38400)
Processor
ChipsetMajor OEM Samsung, Micron, Hynix
Memory
CAS latencyCL40
ECC supportYes
Buffer typeRegistered
Heat spreaderNo
Memory rank2Rx4
Power
Rated voltage1.1 V
Compatibility
Compatible systemsTHINKMATE
In the box
Included items512 GB (8 × 64 GB) DDR5 4800 MHz PC5-38400 2Rx4 1.1 V CL40 288-PIN ECC RDIMM Registered Server Memory KIT

Questions about this item

Does the CL40 latency become a bottleneck when the server runs continuous compute workloads?

The modules operate at DDR5 4800 with CL40 timing; latency is fixed by the JEDEC specification for this speed grade and does not change under sustained load.

How much heat do the eight 64 GB registered DIMMs generate when the system runs 24/7?

Each module runs at 1.1 V without a heat spreader; total power draw scales with all 512 GB active, so adequate chassis airflow is required to keep temperatures in spec.

What server configuration is required before this 8-module kit can be installed?

The kit is validated for the THINKMATE GPX-XH8-22S4-8GPU platform and requires a motherboard with 288-pin DDR5 RDIMM slots that support registered ECC memory.

More in Model-Specific Memory Upgrades

Similar items