Hi, I want a laptop to be able to do my NeuroImaging Analyses, I have acess to a work desktop computer if I need to do more powerfull analyses but would still like to be able tu run Group analyses with SPM (fMRI, VBM, resting state with Conn toolbox) as well as basic tractography with diffusion toolkit and tract reconstructions with Trackvis for example.

I see 2 laptop that I like one has 8GB RAM but with Nvidia MX150 (2GB)
the other one has 16Gb RAM but intergrated Intel UHD Graphics 620

What is more imprtant,more RAM or the beter Graphics Card?

(they both have a i7 procesor 8th Gen, 512 SSD)

Or neither is good and I really need both (16GB RAM with a dedicated Nvidia?

Thank you

3rd Option, What bout 16GB RAM + Nvidia MX150 card but with i5 8250U Processor, would that be better than the 2 previous options? Does the i7 8550 really makes a noticable difference over the i58250?

3rd Option, What bout 16GB RAM + Nvidia MX150 card but with i5 8250U Processor, would that be better than the 2 previous options? Does the i7 8550 really makes a noticable difference over the i58250?

I would go for the second option. You should make sure you could eventually upgrade your graphic card. Having a i7 will save you time in processing tasks. You may ask a configuration with 2 slots of 16 GB (one empty) for future upgrade.

Natale Quartuccio

I'm not sure it's possible to upgrade the Graphoics card on those lightweight laptps.
Also I read that the 8th generation i5 was only marginally faster than the i7 (about 10% dependig on the task) so I'm wondering if I would really see the difference... my choices were Intel Core i7-8550U or i5-8250U)

Originally posted by Elise Barbeau:
Hello !

With the emergence of DeepLearning and GPU accelerated software I would definitely take the third option.
Or at least one with a "real" GPU.

You'll not be able to update the GPU unless you take something like an Alienware (wich is more a transportable PC than a laptop)

Having a faster CPU will only provide a little performance improvement which is not necessary unless you do real time image processing.
Having a real GPU will provide a huge performance up compare to the intel integrated graphics as soon as the app use the GPU, which will be more and more common.

Best,

Odelin Charron

Thank you! The card is the NVIDIA MX150 with 2Gb which is not the best but it's still better than the integretaed Intel one...

Originally posted by Elise Barbeau:
Hello

From my experience neither will be good for larger datasets using SPM and CONN
Unfortunately SPM mostly runs as single thread so the processor performance and clock really matters.
Additionality CONN needs quite good amount of RAM when running multiple subject analysis.
For Trackvis you really need external graphic card.

For neuroimaging I would recomend fully customizable laptops.
Those laptops can be equipped even with 128GB RAM 5xSSDs and Two Desktop Graphic Cards and Desktop Processor up to 5Ghz Intel i9 9900k
There are several manufacturers in Canada US and Europe so you can choose the configuration which will suit your needs without sacrificing anything.

Best
Przemyslaw Podgorski

I know I could get a better laptop but now I'm looking for a lightweight and thin one (not a heavy gaming laptop) with which I can do at least some of my analyses on (I have access to a stronger desktop if needed), I think the specs I found is as good as it gets for those lightweight laptops, but trying to get the best configuration...

Based on what your expected use of the laptop is. I'd skip the dedicated graphics card, as you already have a powerful desktop available. Try to get the highest amount of RAM and clock speed and you will be fine, from personal experience, you would also be able to skip any issues with thermal performance as you're looking for a lightweight laptop, which I would guess means at most one fan and not a proper heatsink.

As a side note, can't you just remotely access your desktop and run the analyses on that?

Best Regards
Emile Berg

"Neuroimaging data" is extremely vague and nonspecific. What kind of imaging, how big are the files, and what kind of processing will you be doing on it? Are there GPU implementations of the algorithms you plan on using, and if not, do you plan on writing them? What kind of redundancies do you want in place for the storage? Whatever it is, I would recommend not getting SSD hard drives - they'll drive up the cost to insane levels for no good reason if you need any kind of big storage capacity.