0
$\begingroup$

Hope all is well! I am planning to buy a laptop (Windows) and looking for your help to decide the specs that we need to speed up the numerical simulations using Mathematica. The Mathemtica code I used usually need a lot of memory to find the results as it involve to calculate a huge matrix such that each entry requiring to calculate many numerical integrations. I am thinking of a iCore 7 processor, 8GB Ram and 512 GB SSD storage. 4 CPUs are enough? For the processor, possibly a turbo speed of 3.60GHz for Intel Xeon 5118? Thanks in advance!

$\endgroup$
  • 2
    $\begingroup$ A new iCore 7 with only 8GB RAM! You're kidding. Mathematica wants RAM. A lot. (I can work quite neatly with 16 GB.) $\endgroup$ – Henrik Schumacher Apr 25 at 17:44
  • $\begingroup$ That’s why I am consulting you guys! I am not familiar with the computer stuff. $\endgroup$ – Mutaz Apr 25 at 18:29
  • $\begingroup$ Erm. I do not think that you can get laptop with a Xeon 5118. That's a server processor. $\endgroup$ – Henrik Schumacher Apr 25 at 22:08
  • 1
    $\begingroup$ @Mutaz if you’re not overly concerned with having 8 cores, and you feel 6 might cut it, I personally use a Powerspec 1520 as my mobile workstation. It’s got a decent 6 core intel processor and the ram is upgradeable to 64gb (2x32gb of what is essentially MacBook ram). If that isn’t enough it’s got an 8gb rtx 2700 as well. I find it suitable enough ;) I haven’t been able to cause a kernel dump/reset from overfilling the usable ram yet. I’m not the most efficient programmer, of course, but I filled up 16 gb pretty regularly doing my work. $\endgroup$ – CA Trevillian Apr 27 at 5:56
1
$\begingroup$

Mathematica as far as I've understood can do parallel processing...and does support some CUDA things.

To use Mathematica's built-in GPU computing capabilities, you will need a dual-precision graphics card that supports OpenCL or CUDA, such as many cards from NVIDIA, AMD and others.

With that said however, the majority of things you'll do is probably singlecore stuff \

For large symbolic calculations, on my 16GB machine, I've actually run out of ram. So I can attest from personal experience, more ram is better, ideally lots and fast.

So basically, if your budget can handle it, the fastest CPU (probably 4c at the minimum...[can one even buy less than 4c nowadays?]) and ram you can get.

A dedicated video card would be nice, but not hugely required, depending on what you want to do.

If you check out the recommended hardware page, you'll see requirements are quite modest at most.

Depending on what you do, if you solve an equation in 0.0001 or 0.00001 seconds...you arn't going to notice a difference.

I've been running my MMA on a macbook pro from 2013 for a while now, it still runs buttery smooth in version 12 and I've rarely done anything to max out ram or slow the software down...

My specs being:

  • i5 2.4ghz
  • 16gb 1600mhz ddr3
  • intel iris

If you can do better or equal to that, I suspect you'll be satisfied.

| improve this answer | |
$\endgroup$
  • $\begingroup$ Much appreciated! $\endgroup$ – Mutaz Apr 25 at 19:54

Not the answer you're looking for? Browse other questions tagged or ask your own question.