Skip to content
SUPERCOMPUTERS

Numerical computing · MathWorks

Workstations for MATLAB

Numerical work where built-in maths uses every core, your own loops use one, and the graphics card helps only in specific cases.

Task by task

What MATLAB actually needs

1

Scripts and loops

Needs: One fast core

Code you write yourself runs on a single core unless you parallelise it.

2

Built-in maths and parfor

Needs: Cores and memory bandwidth

Matrix maths uses every core automatically, and parfor spreads loops across them.

3

GPU computing

Needs: An NVIDIA card, for single precision

gpuArray moves work onto NVIDIA cards and is very fast in single precision.

The same advice applies to Simulink, Python with NumPy and SciPy and Julia.

Hardware at a glance

What to buy for MATLAB

ProcessorHigh clock for serial code; 24 or more cores if you use parfor or large matrix work.
MemoryAt least three times your largest dataset. Error-correcting memory for long runs.
GraphicsAn NVIDIA card if you use gpuArray in single precision. Otherwise a modest one.
StorageNVMe for datasets and results.

Our picks

Machines we recommend for MATLAB

Recommended machines compared
Specification
Our takeFor coursework, Simulink and research code that mostly runs in one thread. The fastest per-core machine at the lowest rung.For parfor, large matrices and runs measured in hours. 24 cores and 128GB of error-correcting memory.For the largest datasets and long parallel runs. 32 cores, 256GB and a 48GB card with error correction.
ProcessorAMD 12 Core ProcessorAMD 24 Core ProcessorAMD 32 Core Processor
GraphicsPalit 16 GB Vram graphics cardNvidia 24 GB Vram graphics cardPNY 48 GB Vram graphics card
Memory32GB High speed memory128GB High speed memory256GB High speed memory
Storage2TB total3TB total3TB total
FormTower, desk-sideTower, desk-sideTower, desk-side
Actions

Questions

MATLAB hardware questions

Does MATLAB use multiple cores?

Built-in functions such as matrix operations do, automatically. Your own loops do not, unless you rewrite them with parfor from the Parallel Computing Toolbox.

Does MATLAB need a graphics card?

Only for GPU computing with gpuArray or deep learning. It must be NVIDIA, and it is strongest in single precision.

Why is error-correcting memory recommended?

On runs that take hours or days, a single flipped bit can change a result without any error message. ECC memory detects and corrects it.

Why buy from us

91checks

Checked before it is quoted

Fit, power, cooling and memory, on every build.

72hours

Burned in before it ships

Full load for three days, so faults show up here.

1working day

An engineer replies

With a specification and a quote for your work.

1number to call

Supported in Pakistan

From Islamabad, delivered nationwide, with warranty handled here.

Other scientific computation software

All software

Want an engineer to size it for your MATLAB work?

Tell us the size of your largest model. An engineer replies within one working day.