x170sm-G for max compute power?

I am looking at this guy for my needs of maximum computing power in a laptop. I do not care about gaming - only scientific computing, which means maximum number of cores at maximum speed. From reading other threads I understand that the i9-10900k may use much more than the 125W it is listed at when doing intense calculations in parallel. How can I find out what this build will allow as far as, say all 10 cores running at once? Will it allow them all to reach maximum potential? If not, then would a lesser processor be better?

I am interested in the most cores running at maximum speed for sustained computing. So perhaps 8 cores running at 4.7Ghz would beat 10 cores running at 3.5Ghz. Can someone help me out here?

If you had to spend <5k on a laptop and you only cared about maximum computing power - not graphics so much - what would you get?

https://xoticpc.com/collections/ultimate-series-gaming-laptops/products/sager-np9670-clevo-x170sm-g

 

XMG Apex 15 with 16 core Ryzen 3950X will comfortably beat 10900K in compute benchmarks.

 

Even neutered at 65W?? Does someone put a non limited one in a laptop?

 

16 cores are 16 cores and apparently boosts to more than 65W in performance mode, 7600 CB20 vs 9200 ish at full power. Reviews are ez to google.

User's lounge http://forum.notebookreview.com/thr...f1-owners-lounge.832568/page-63#post-11035328

 

Thanks, I saw that one. The trouble is there so much conflicting information on that long thread. Nobody seems to answer the question I need answered which is what clock speed will all the cores run at simultaneously for sustained periods in that particular laptop? That's the only way I can think of to compare two laptops. For example if I have one with 10 cores and they all run at 4 gigahertz then that's approx. the same as one that has 20 cores all running at 2 gigahertz. But if the 20 core machine runs at 1 gigahertz then it's is not as good for computing.

Thanks for your input nonetheless.

 

For example, here the 10900k kills the AMD, but I think there is a mix of desktop and laptop so I am not able to parse them.

 

You have to know as well, that the XMG Apex 15 maximum GPU is RTX 2070, so if your scientific applications needs the GPU power then you have to go with the RTX 2080 Super with the 10900k.

 

Don't know about the clocks (surely they will be in the low side if all 16 cores are loaded with 65W), but benchmark results look convincing, and the price is phenomenal compared to 10th gen Intel. Why do you care about clock speeds? Is your application latency sensitive?

 

Those aggregate comparisons are not that useful, better to look at specific benchmark results, for example on notebookcheck.net. The 10900K will do better in gaming benchmarks, however, it will get beat in one specific application: mutithreaded/multicore CPU performance (e.g. CB20) which is what the OP cares about.

 

He may have already said he doesn't care about that and the 2070 is the sweet spot anyway. With the Apex he can also select a cheaper GPU, namely the 2060, to avoid paying a premium for a fairly marginal improvement he doesn't need.

 

Not mocking at anybody... I feel that the market needs a portable EPYC system for EPIC Computing Calculations for enthusiasts like OP!
Their devotion in Computing findings will make technology take a leap to new heights!

 

I thought the one with 1660Ti is the lowest of the NH5#A variant? SAGER offers one for $1100 with the base configuration of 3300.

 

Apologies, my comment mistakenly referred to Appleware A51M R2 which is offered with 2070S at a minimum, for 3100 euros, 50% more than the 2060 3950x Apex.

 

I would rather AMD mounted a strong attack on the portable laptop sector, which, in addition to strong performance, means an Optimus-like solution and ability to operate at a few W when idle.

 

Am Epic based system would make the old P570WM look small and would make it seem like it had a good battery life.

 

Isn't clock speed a measure of how fast the CPU runs? My first machine had a PIII 512 MHz and it's was slooow compared to my 4 GHz 6700K.

 

I mainly use CFD applications and custom Fortran codes written with MPI, none of which run quickly or at all on GPUs.

 

Under full computational load and given identical architecture and number of cores: yes, there are a crude benchmark, but they don't capture other factors contributing to performance. In some cases, the clocks can be very misleading - for example, I can make all cores in my current laptop run at 3.9GHz, while the CPU is idle and consumes 2W.

 

Lenovo is gradually annually increasing the model option of AMD based laptops for the RED "SOCs" followers. However to date, we're still not seeing a system design for workstation P-series. The only addition this year of AMD to their office product profile is the L series.
Others? Well... in the gaming field, INTEL system still reigns in the position of offering "bang for buck" performance. Hence, conservative buyers will still tend to believe INTEL based systems gives best gaming experience...
The Ryzen 4000H series is a turning point in 2020 to gain more attention on gaming performance, however, without a very effective cooling design like what Lenovo had achieved with their Legion 5 lappies, temperature still climbs to "worrying state" to owners of non-Legion buyers of other makes. This gives a very negative impression to interested buyers of AMD based products even though they may not be buying the gaming machines.

Bad impression gives Bad influences.

Yes. As we goes more mobile, laptop sales is perhaps one of the most effective route to gain more trusts on AMD products. However it seems that OEMers are still on the tab by INTEL's deal not to go freely fully to produce AMD based products, hence lesser choice and enthusiasm from the OEMers. On the consumers' end, the trust on INTEL magnificent power in computing is still strong... You can feel the warmth in this thread.

On AMD. IMO, they should pay attention now as more apps and softwares are designed towards ARM devices. And AMD should be in the work of reverse engineer their ARM certified Opteron CPU to see if they can develop a ARM CPU not only for mobile phones but also mobile computing.
After the arrival of Apple's ARM laptops in coming days, the new route in computing is paved, and it's up to AMD ARM product's performance and acceptance to shine in that path.
The x86 route is now slowly populated by the RED Muscle-Cores, however they're also "perceived to be power guzzlers", that those who appreciates are growing gradually in small numbers... All i hope, is that consumers will realize one day that they're paying more per year with Blue Muscle-Cores yet the performance takes a very gradual improvements.

 

OK, but isn't there some relationship between wattage and Max speed? As in, wouldn't that machine run faster for longer under more load with 105 watts instead of 65? I assume temperatures kept constant.

 

In this case faster - yes, for longer - not necessarily, under more load - the user controls the load. Generally there are diminishing returns in terms of power supplied to a given chip, and power comparisons across different types of CPUs are almost meaningless, which is why a 16 core CPU running at 65W, can outperform a 10 core CPU running at 100W.

 

Thanks for everyone's answers. I have just a few more questions.

1. Is it hard to partition a hd that comes with windows to dual boot some version of Linux?
2. Are there certain programs or hardware that won't work with AMD?
3. Is this equivalent to that English maker?

https://www.hidevolution.com/evoc-h...-3900-ryzen-9-3950x-gtx-1660-ti.html?___SID=U

 

https://system76.com/laptops/serw12/configure

 

My answers to you in quote.

 

Re 2.: this reddit thread sheds some light:

https://amp.reddit.com/r/Amd/comments/a2jk2q/what_are_the_honest_limitationsissues_with_amd/

There should not be any hard incompatibilities per se.

One thing they mention is the lack of AVX-512 support:

https://en.m.wikipedia.org/wiki/AVX-512

However, this is implemented in Intel Xeon CPUs rather than the 10900K.

https://ark.intel.com/content/www/u...0900k-processor-20m-cache-up-to-5-30-ghz.html

They also mention some difference in AVX instruction timing.

One thing that could happen is that software heavily optimised for Intel CPUs performs less efficiently on an AMD processor. Ideally, you should investigate this with the vendor of your core numerical libraries.

 

Is anyone interested in an RTX 2080 Super?
It is as good as new.
Greetz

 

If it's MXM, yes I'm interested .

 

Sure, it of course an MXM card.

 

Would you PM me all details along with the price,

 

It's more like the current desktop Intel mainstream arch is do old AVX 512 was not around yet, the next gen cores should have it from Intel.