alkakxmmiq

The specification is out for almost 2 years but not any products. Here on Ars Technica (mid 2019 article) they said: So according to that article they should already be coming out. But everything seems to be getting delayed nowadays, so...

I haven't heard anything about USB4 rollout lately. Have you? Frankly Thunderbolt is the last killer feature of Intel laptops remaining so when USB4 becomes available on AMD laptops, I'm buying one.

nVidia is stopping production of the 2000 series GPUs so 3000 series is coming in the next months. When do you think we will start to see laptops with the 3000 series GPUs? Based on the fact that AFAIK only the production of higher end cards is stopping, the high end (and high power) 3000 series cards are probably going to be released first but I don't know how much time it will take nVidia to release the lower end cards and how much time from their release it will take laptop manufacturers to build laptops with them. What are your opinions?

I'd like to try some fluid dynamics or other physics simulations to put my education to use ? Although you might be right that for getting into GPU computing I could try it out on something less expensive. I've been thinking about buying a Jetson Nano to try it out first...

Light gaming but I'm considering it mostly for its unmatched double precision (fp64) performance since the card that can beat it at that is Titan V at 4-5 times the cost. Also considering trying out ML workloads where nVidia shines more with Tensor cores and superior software stack but Radeon VII is not bad considering the memory size and bandwidth.

Hi! I was hoping that the release of nVidia's SUPER cards would cause the Radeon VII's price to drop but it seems AMD is killing it instead. Nevertheless, I would like to buy one if the price drops but I don't know whether it will. Do you think the prices of Radeon VII will drop at least for clearance of the remaining stock? I haven't seen any downward trend in pricing recently as I thought would have to happen. Also what do you think about driver support? Will it continue as Radeon VII is basically a cut down server-grade GPU or will it end soon?

But it has only 1/32 double precision performance compared to its single precision (32bit float) one. The Titans and such have modes where you can get 1/3 (or 1/2 in the case of Titan V) double precision performance compared to the regular precision. I'm on odds with the 1080ti for that reason ? There are so many variables, there is no clear winner and I'm a really indecisive person ?

Thanks for the experience! I am thinking more about the Titan Blacks in this category since while searching local second hand deals I have deals which are about 40 % more expensive than 660s but it has incomparable performance on the double precision calculations which is crucial for finite element method calculations, n-body simulations and such. On ML I think one would rather have half-precision with greater speeds, that's why I'm asking for someone's experience on the RTX cards. Anyway thanks for the input, it's great to have some real-world data so I can get a feel for how long it will run these tasks.

I hope there will be plenty available now after Christmas ? So I'm keeping my eye on them. It's great that top of the line 1080ti is not that more expensive than a new 2070. The real question is: Can I live without raytracing? ?

Hi! I am thinking about buying a new PC but can't decide which GPU is the best bang for buck for me. I dream of a Titan V which is hands down the best for my purposes (tensor cores + 1/2 fp64 FLOPS compared to fp32 FLOPS is a killer feature for computer modelling in physics since normally nVidia cards have 1/32; sadly they disabled ECC) but it is probably unwise to spend that much money since I'm just starting. So I'm going mainly for bang for buck. The main contestants are RTX 2070 or second hand either 1080ti, Titan Z, titan black or perhaps some others. Titan Z has great fp64 performance, not bad fp32 performance but has a really high TDP and an older architecture and can be purchased for the lower end of the price of the second hand 1080ti. Titan Black is really cheap and has pretty good fp64 performance, otherwise it's quite old. 1080ti is all round great (memory, bandwidth, fp32 performance) but does not have Tensor cores (does anyone have experience with how tensor cores boost performance in ML compared to normal parallelization?), not that much performance on fp64 and has higher TDP than 2070. rtx 2070 does not have the most memory, memory bandwidth or fp32 performance, but all of it is decent. fp64 performance is not that great but it has Tensor cores (again, does anyone have experience with them in ML applications?), great fp16 performance (idk how the older ones fare on fp16), DLSS and HW ray tracing for games or graphics, new architecture and good performance/TDP ratio. other considerations: I've looked at eBay and they have used tesla k80s for about 650 USD. Is that possible? If seems like a great deal since k80 has ECC, great fp64 performance and lots of memory. Sadly no fp16 performance boost like on Volta and no tensor cores but the ECC is still really valuable for simulations (not that much for ML afaik), also I would have to use something else for gaming and other kinds of work since it does not have a display output (is it possible to wing it somehow?). tl;dr: Do you have experience with workloads mentioned in the title? Which GPUs would you recommend for them taking into account bang for buck performance.

WOW! I just went to the topic and the post is there just unposted! So I just posted it. All hail the LTT forums, they are the best!

Budget: Around 25k CZK (location: Czech republic) ~ 1100 USD ~ 965 EUR at the moment. Why upgrade: I currently have only an old Thinkpad notebook with internal graphics and I can do most of the stuff but I can't do some computationaly intensive tasks I would like to be able to do. Time: Not really in a hurry so I would also like some recommendations if I should buy now (perhaps snatch something juicy second hand after Xmas?) or wait for the next year (3rd gen Ryzens? Prices of some components could go down?) Aims: All sorts of stuff. 1) Gaming: Minecraft, Dota 2, CS:GO etc. Don't have time now for the AAA stuff but when I will I would like to be able to play things like Kingdom Come or Dark Souls III. 2) Machine learning: I plan on tinkering with neural networks a bit since it is becoming an essential stuff for modern computing and has lots of crazy cool applications (eg. this is insane https://arxiv.org/pdf/1803.04779.pdf and a popular article about it https://www.quantamagazine.org/machine-learnings-amazing-ability-to-predict-chaos-20180418/ ) 3) Computational fluid dynamics: I have a degree in physics where I've done some fluid dynamics theory but never ventured to the practical side that much and I would also like to do that. For that I need good double precision (fp64) performance. 4) Virtual machines and server stuff: Would like to play with Xen, try running some servers and such. 5) Web browsing: One need lots of RAM for this one nowadays ?. Monitors: Now I have just one entry level PLS 1080p 60 Hz monitor (with freesync) and I seem to be fine with that. Maybe in the future another one will join it. I haven't seen 4k monitors IRL so I can't judge the experience but 1080p seems enough for me for now although I would like to have room for improvement in the future. Peripherals and already owned parts: I have a decent mechanical keyboard (great experience) and a good mouse and decent speakers. I already have a 120 GB and a 500 GB SATA SSDs so I would like to use them (maybe I will move to NVMe at some point but I would have to be convinced it is a really big performance jump over SATA SSDs and I seem to be happy with their performance for now). I also own a 4 TB HDD which I would use for storage so hard drives are already solved. Questions and considerations: Security: Do you think it is worse to have IME or PSP in your system? Intel vs. AMD choice regarding CPU is less about the models themselves for me (both have really good CPUs) but more about IME vs PSP. If someone knows about that stuff (tried me_cleaner or so) please post your considerations. GPU: I am currently considering RTX 2070, gtx 1080 ti second hand or Titan Black second hand. Which one would you recommend? Perhaps a different one? RTX 2070 pros: tensor cores for machine learning - how much of a performance boost is it compared to running it on 1080ti or titan black? 16bit number performance - also for machine learning DLSS and ray tracing cores for gaming new architecture and interfaces (CUDA 7.5 some slight changes to the shaders) ?better performance/power ratio? (is this true?) highest effective memory clock cons: less fp32 performance than 1080ti (less shaders) no SLI less fp64 performance than Titan black (1/32 of its fp32 performance) gtx1080ti pros: best fp32 performance better fp64 performance than 2070 same price for a good card (even hybrid cooling) as for 2070 new entry level card 11GB memory highest memory bandwidth (has 352bit memory bus) SLI cons: higher TDP than 2070 older architecture no tensor or RT cores, no DLSS no 16bit mode for ML less fp64 performance than Titan black (1/32 of its fp32 performance) Titan black pros: super low price - Can buy 3-4 second hand Titan Blacks for the cost of 1 1080ti or 2070 bonkers fp64 performance (has a mode where its fp64 performance is 1/3 of its fp32 performance instead of 1/32; 5.5x better than 1080ti and 8x of 2070) SLI cons: worst performance of the 3 on everything apart from fp64 less memory worst memory bandwidth old architecture higher TDP than 2070 Memory: 2x8GB seems to be good. Model depending on the MB MB: depends on the CPU PSU: I would like a recomendation on this one. case and looks: I don't fancy cool looks but rather performance so any case with good thermals will do, no Xmas lighting needed. I would rather invest the few bucks otherwise spent on a nice looking case to other components.

I will do that but it didn't occur to me at the moment that it could happen. Ok. I will post a brief version now and see where it gets me.

Yeah. Best to invest the 1100 USD into a decent PC. I'm thinking about building one for a long time but I'm really indecisive and new tech always seems to be around the corner ? I've almost posted on this forum about the build. I had written a nice post but I had it opened in browser overnight and when I hit send it logged me out and the post got lost so I am a bit angry now.