91–100 of 108 posts
Playing around with 96x avx512 sounds fun
Pardon my ignorance, I hear about "fun with avx512" so often yet still have 0 intuition about why its fun? What do people do with that instruction that's so different to all the other stuff?
Has anyone done a perf / watt scatterplot of the current AMD and Apple offerings? I'd guess this wins in absolute performance (due to the TDP), but I've heard AMD is pretty competitive on energy efficiency too these days. I'm typing this on an M2, which is simultaneously cooler and faster than my old Intel laptop. Wondering if it's time to upgrade my decade-old Intel Linux desktop to an AMD yet.
> I'd guess this wins in absolute performance No need to guess: An M2 Ultra will get around 27,000 in Cinebench R23 while this Threadripper will get around 100,000. The M2 Ultra has a TDP around 90 watts (~ish according to a couple reviews) and this Threadripper has a TDP of 350 watts.
The stock configurations for a lot of desktop chips is to get the maximum performance at the expense of power, especially for things like these Threadripper series which are explicitly aimed at maximum performance. Often even drastically reducing power usage only gives modest performance penalties.
Has anyone done a perf / watt scatterplot of the current AMD and Apple offerings? I'd guess this wins in absolute performance (due to the TDP), but I've heard AMD is pretty competitive on energy efficiency too these days. I'm typing this on an M2, which is simultaneously cooler and faster than my old Intel laptop. Wondering if it's time to upgrade my decade-old Intel Linux desktop to an AMD yet.
For my m1, opening a vm or just plugging in a wireless mouse dongle is all it takes to shave several hours off the lifespan, like from 20 hours to 10.
For the amd laptop, anything waking up the nvidia gpu will do the same, like 10 hours to 3.
When you're talking such low amounts of power, the networking card, sound card, display efficiency, ssd efficiency, and whether the system chooses to run the fan or let the cpu stay warm starts to matter a lot since just like the cpu these are things that are running at the same time. So I guess a fair test would have to use the same ssd and have the monitor off, using an external monitor. Then you're just stuck with the efficiency of the parts of the motherboard you can't choose.
This will be perfect for my Plex server.
Wouldn't a cheap GPU be better for transcoding tasks?
Earlier quoted context omitted.
> I'd guess this wins in absolute performance No need to guess: An M2 Ultra will get around 27,000 in Cinebench R23 while this Threadripper will get around 100,000. The M2 Ultra has a TDP around 90 watts (~ish according to a couple reviews) and this Threadripper has a TDP of 350 watts.
If you underclock the Threadripper to get a score of ~27k, then how much power will it use? I wouldn't be surprised if it was roughly equal to 90W, or even lower. The stock configurations for a lot of desktop chips is to get the maximum performance at the expense of power, especially for things like these Threadripper series which are explicitly aimed at maximum performance. Often even drastically reducing power usag…
I also run an RTX 3900 GPU at 60% power for 80-90% performance. It's a noticeable decrease in requisite cooling.
Earlier quoted context omitted.
Actually, no - their thing for the past few years has been to build out the most comprehensive test bench across all PC hardware components. "LTT Labs." They are aiming to become the authority on hardware performance and reliability, and have invested heavily in the tools and industry experts required to do that. Linus doesn't do any testing himself; he is just bankrolling it with the proceeds from his main channel a…
I don't know when/if that change you talk about happened, as I banned LTT from my feeds years ago for their (subjectively) annoying style of presentation, but latest drama isn't even year old and highlighted so many technical issues with LTT I have no intention of trying them again: [0]: The Problem with Linus Tech Tips: Accuracy, Ethics, & Responsibility [0]: https://www.youtube.com/watch?v=FGW3TPytTjc
I am just wondering if it would be better and cheaper to buy an Threadripper 7980X 64 Core for Server than EPYC. You get higher single core performance. Specifically for those workload where bandwidth isn't an issue. Something like StackExchange [1] could fit all their 9 server into 1 ( or 2 with one for redundancy ) without any degraded performance. [1] https://stackexchange.com/performance
Nobody believes me, which says a lot about both the hardware and software industries.
I am just wondering if it would be better and cheaper to buy an Threadripper 7980X 64 Core for Server than EPYC. You get higher single core performance. Specifically for those workload where bandwidth isn't an issue. Something like StackExchange [1] could fit all their 9 server into 1 ( or 2 with one for redundancy ) without any degraded performance. [1] https://stackexchange.com/performance
I keep trying to make the point that one or two servers could now put out the traffic of many top 100 websites, assuming they apps are written reasonably efficiently (e.g.: with C# or Rust, not Ruby). Nobody believes me, which says a lot about both the hardware and software industries.
I would be happy if Ruby Rails could do it in 2x the rendering time i.e slower with 2x Resources / Server. That is combined 4x difference. Unfortunately even with JIT we are not there yet.
What use cases is this CPU ideal for? It's not a GPU, so gaming, ML model training, and crypto mining are out.
These things can chew through compiles so fast I/O to disk will be the bottleneck, rather than actual compilation time. Great for Rust hackers.
What use cases is this CPU ideal for? It's not a GPU, so gaming, ML model training, and crypto mining are out.