SC88

SC88 — 48-Core, 1 TB CFD HPC Node for Large, Memory-Bound Solves in ANSYS Fluent

The SC88 node is the top of our range: a flagship built for the biggest, most memory-hungry CFD cases. With 48 cores and 96 threads across two Intel® Xeon® Platinum 8160 processors and a full 1 TB of RAM, it is made to load and solve very large meshes in ANSYS Fluent — cases that simply will not fit on a smaller node.

Where our SC6–SC10 line trades memory for a keen price on compact cases, SC88 is the opposite end: maximum memory and a modern Skylake-SP core count for the simulations that push hardware to its limit.

Specifications

The strength of SC88 is headroom: 1 TB of memory feeds 48 modern cores, so large and physically complex cases both fit and run — no compromise between mesh size and solve speed.

What this node is built for

SC88 is the large-case, high-memory tier. The 1 TB of RAM sets the ceiling on mesh size, and the 48 Skylake-SP cores with AVX-512 and 6-channel memory keep those large meshes solving at pace. It suits:

• Very large meshes — hundreds of millions of cells for steady single-phase cases, with room to spare that smaller nodes cannot offer.

• Memory-heavy physics — multiphase (VOF, Eulerian), combustion and detailed reaction chemistry, DPM with many parcels, radiation, and moving/overset meshes, all of which push memory well beyond the simple 1 GB-per-million-cells baseline.

• Large transient (LES/DES) runs — high cell counts plus fine time-stepping, where both memory and sustained core throughput matter.

• High-fidelity production cases — final validation and design-of-record runs where you refuse to coarsen the mesh to make it fit.

• Memory-bound cases — where the bottleneck is RAM, not core count, and a smaller node would fail to allocate or fall back to disk.

With 48 cores at 2.1 GHz (turbo to 3.7 GHz) and six memory channels per CPU, Fluent partitions the mesh across all cores and streams data through high memory bandwidth each iteration — so even a very large case advances steadily rather than starving for memory.

Where SC88 fits — and when a smaller node will do

In CFD, memory sets the largest mesh you can run — roughly 1 GB of RAM per million cells for a typical steady, single-phase case, and considerably more for complex physics. SC88's 1 TB is our largest budget, matched to the meshes and models that overwhelm everything below it.

If your case is smaller, you may not need this tier. On the same rental model you can step down to:

• SC10 (128 GB), SC9 (96 GB), SC8 (64 GB) for large-to-moderate meshes,

• SC7 (32 GB) for moderate meshes,

• SC6 (16 GB) for smaller cases you want solved fast at a keen price.

Tell us your cell count and physics, and we will match you to the right tier — SC88 when you need the ceiling, a leaner node when you do not.

Fast SSD storage

The 2 TB SATA SSD gives large case and data files room to live and responsive read/write for loading, writing, and autosaves. On big transient runs that write frequently, that keeps a long solve from stalling on slow disk I/O.

How parallel solving works on this node

Running Fluent in parallel, the solver partitions your mesh across all 48 cores (using algorithms such as METIS), each core solving its share of cells and exchanging boundary data each iteration via MPI. On a single node, that communication runs through fast on-board channels rather than a network, keeping parallel overhead low. Combined with AVX-512 vector units and six DDR4-2666 channels per CPU for high memory bandwidth, a well-partitioned large case is solved efficiently across all 48 cores.

Why rent instead of buy

A 1 TB machine is the hardest tier to justify buying — it sits idle between big jobs. Renting SC88 gives you that ceiling exactly when a large case demands it, and nothing to maintain the rest of the time.

How it works

1. Tell us about your case — mesh size, physics, and timeline. We confirm SC88 fits (or recommend a leaner tier if you don't need the memory).

2. Upload your case files securely.

3. We run it on the node with ANSYS Fluent.

4. You receive your results, ready for post-processing.

Frequently asked questions

How is SC88 different from the SC6–SC10 line? Those nodes share the older 20-core E5-2670 v2 platform and range from 16 GB to 128 GB. SC88 is a different, newer platform entirely — two Xeon Platinum 8160 (Skylake-SP) CPUs for 48 cores / 96 threads, 1 TB of RAM, AVX-512, and 6-channel memory. Choose SC88 when your case needs far more memory, cores, or per-core throughput than the SC line provides.

How large a mesh can it handle? With 1 TB of RAM, plan for hundreds of millions of cells for simpler steady single-phase cases, and proportionally fewer for transient, multiphase, reacting, or radiative physics (about 1 GB of RAM per million cells is a starting estimate, and complex models use more). Send us your case for a precise answer.

Do I need all 1 TB? Only if your case does. If it fits comfortably in less memory, a smaller node will be more cost-effective — we'll tell you honestly which tier matches your simulation.

Is my data kept confidential? Yes.

Get started with SC88

If you run large or memory-heavy CFD cases that smaller nodes can't hold, SC88's 48 cores and 1 TB of RAM are built for exactly that.

Contact MR CFD to run your ANSYS Fluent case on SC88 — and we'll help you confirm it's the right tier for your simulation.