Hardly a mistake; whether you bought Intel's 11K, 12K or AMD's 5000 series CPU's, they're all stupidly fast, I can't see how you could make a mistake at all.Wow, Intel 12900K is upping desktop performance even with DDR4 memory. I think it was a mistake that I recently built a 5950X system. Sure, it's fast, but I should have waited for Alder Lake.
Single-core performance is one of the biggest factors in gaming and most of the other applications I run (e.g., PhotoShop, SolidWorks, etc.). For the same amount of money (or less) I could have had a system with significantly faster single-core performance. Also, I've some reliability problems (lockups and reboots) in my 5950x system that I never had before in previous (Intel) builds.Hardly a mistake; whether you bought Intel's 11K, 12K or AMD's 5000 series CPU's, they're all stupidly fast, I can't see how you could make a mistake at all.
Single-core performance is one of the biggest factors in gaming and most of the other applications I run (e.g., PhotoShop, SolidWorks, etc.). For the same amount of money (or less) I could have had a system with significantly faster single-core performance. Also, I've some reliability problems (lockups and reboots) in my 5950x system that I never had before in previous (Intel) builds.
I don't believe that is always true. If money is not the object then someone might choose the fastest possible performer. I provided a link below to the ranked single-core performance scores for most CPUs. Note that the single-core performance of a 5600x is not faster than a 5950x. In fact, the 5800x, 5900x, and 5950x have almost exactly the same single-core performance, so it makes sense to choose a 5950x if you can afford one. And, many of the higher-end Intel CPUs have higher single-core clock limits, so you usually will get faster single-core performance from CPUs with more cores.Single core performance ? you got the 16 cores chip , people who want single core performance dont get more cores ...
Anantech has been going downhill for some years now.@ikjadoon
Anadtech used CL22 DDR4 for some reason, looked like they wanted to have bad results on that front. CL16 or even 18 (which also comes close) is available at very good prices and perform near or better than DDR5 4800 sets.
I haven't even seen CL22 DDR4 sets available for purchase.
Not true. Especially with Intel, the higher in the stack you go, the higher the chip will boost. 10 core 109000k has low core boost to 5.2Ghz. The fastest quadcore 10 series you can buy is the i3-10320 has a max boost of 4.6Ghz. That additional 600Mhz will make the 10900k a measurably faster single core performer.Single core performance ? you got the 16 cores chip , people who want single core performance dont get more cores ...
I still think you're being a bit nit-picky, but it depends on your values I guess. Once upon a time I used to spend a <Mod Edit>-ton of money on everything 'fastest', but these days I don't care so much as everything is already pretty damn fast.Single-core performance is one of the biggest factors in gaming and most of the other applications I run (e.g., PhotoShop, SolidWorks, etc.). For the same amount of money (or less) I could have had a system with significantly faster single-core performance. Also, I've some reliability problems (lockups and reboots) in my 5950x system that I never had before in previous (Intel) builds.
I kind of agree - do you apply this same philosophy to the cars / bikes / video cards that you buy?I don't believe that is always true. If money is not the object then someone might choose the fastest possible performer.
Not true. Especially with Intel, the higher in the stack you go, the higher the chip will boost. 10 core 109000k has low core boost to 5.2Ghz. The fastest quadcore 10 series you can buy is the i3-10320 has a max boost of 4.6Ghz. That additional 600Mhz will make the 10900k a measurably faster single core performer.
I don't believe that is always true. If money is not the object then someone might choose the fastest possible performer. I provided a link below to the ranked single-core performance scores for most CPUs. Note that the single-core performance of a 5600x is not faster than a 5950x. In fact, the 5800x, 5900x, and 5950x have almost exactly the same single-core performance, so it makes sense to choose a 5950x if you can afford one. And, many of the higher-end Intel CPUs have higher single-core clock limits, so you usually will get faster single-core performance from CPUs with more cores.
https://www.cpubenchmark.net/singleThread.html
Um... what are you talking about? I said I cared about single-core performance. By definition SINGLE CORE performance is measured using one core only, which is often the fastest core. It is not measured using all of the cores simultaneously.you are ignoring the overclocking all together . it is harder to push more cores in OC than lower cores CPU ...
Money is not a concern for me at all because it is for my business, but I hate wasting money when I could have purchased a more performant, and probably more reliable product. In fact, I've already acquired all of the parts I need to build a new 12900K DDR4 system, but I'll keep the 5950x system as a backup.I kind of agree - do you apply this same philosophy to the cars / bikes / video cards that you buy?
In my case money is less of a concern and so the last CPU I bought was a Threadripper. It always makes me chuckle when some people bleat about how they're not for gaming etc. but I don't care... it's my hobby and my money.
Providing I can actually get one - thanks to supply issues etc. my next CPU will be a Threadripper too - and I'll use it to play games. Depending on what happens on Black Friday, I might just build a 5950X platform just for laughs.
Sorry, but if it is for your business, why did you get a 5950X and not a ThreadRipper as it seems you have a use case for the threads and the TR platform has way more PCIe lanes you can actually use for "pro" work? Not even counting the juicy Quad and Octo channel options, instead of the peasant Dual channel memory in mainstream.Money is not a concern for me at all because it is for my business, but I hate wasting money when I could have purchased a more performant, and probably more reliable product. In fact, I've already acquired all of the parts I need to build a new 12900K DDR4 system, but I'll keep the 5950x system as a backup.
Because, as said previously, the performance of applications I run (especially SolidWorks) are tied to single-core performance. None of the ThreadRippers have relatively high single-core performance.Sorry, but if it is for your business, why did you get a 5950X and not a ThreadRipper as it seems you have a use case for the threads and the TR platform has way more PCIe lanes you can actually use for "pro" work? Not even counting the juicy Quad and Octo channel options, instead of the peasant Dual channel memory in mainstream.
Regards.
For example, from the link I quoted above, the 3945WX is the ThreadRipper with the highest single-core performance at 2,714. However, the Intel 12900K's single-core performance is 4,157.Because, as said previously, the performance of applications I run (especially SolidWorks) are tied to single-core performance. None of the ThreadRippers have relatively high single-core performance.
https://www.cpubenchmark.net/singleThread.html
Thanks for the explanation.For example, from the link I quoted above, the 3945WX is the ThreadRipper with the highest single-core performance at 2,714. However, the Intel 12900K's single-core performance is 4,157.
The 12900K is thus about 53% faster at single-threaded calculations than the ThreadRipper with the fastest single-core performance.
What do you mean by mixing workloads? Despite the number of cores, single-core performance is important for 95% of business applications run on a desktop. As I keep saying, I based my purchase decision on single-core performance because the applications I plan to run on THIS computer are sensitive to single-core performance. This 5950x system was my first AMD build. I built it because I figured it would be a while before DDR5 ram for Z690 became readily available. Plus, I figured I could drop in the next-gen AMD CPU when it becomes available.Thanks for the explanation.
I still find strange how you're mixing the workloads, but I at least can see why the 5950X was the best choice for you at the time. if the "wide" loads are not as important, then the 10900K/10850K may have been a better pick for way less money? Well, I guess that's a bit too much nitpicking, but if it's for your business, nothing prevents you from selling the 5950X and getting the new i9, right? It's not like you won't be able to sell the 5950X platform and get back a good chunk of the money back. I guess it'll depend on your accounting, but you can do, right?
Regards.
Does anyone really care?What is the power consumption in gaming?
How you spread the workloads on the PC. Basically the time spend on tasks that actually "give you money" and justify the SP performance over the MT performance.What do you mean by mixing workloads? Despite the number of cores, single-core performance is important for 95% of business applications run on a desktop. As I keep saying, I based my purchase decision on single-core performance because the applications I plan to run on THIS computer are sensitive to single-core performance. This 5950x system was my first AMD build. I built it because I figured it would be a while before DDR5 ram for Z690 became readily available. Plus, I figured I could drop in the next-gen AMD CPU when it becomes available.
However, I was surprised at how performant the 12900K with DDR4 turned out to be. In fact, it seems unlikely that the next-gen AMD CPU will have significantly faster single-core performance, but if it does I will have a system to drop a new CPU into. If not, I will have a 12900K system.
It matters for cooling requirements. Even if you're not going to use 100% of the potential power usage (full PL2), you still need to see how games make the power spike and settle. you can save a lot of money by not going the AIO route. Specially with the i9. The i7 is a bit more manageable, but still can draw 190W, so that is still not trivial to cool if you have workloads you use that require the CPU to go to 100%. If/when games use more than 6 hard threads consistently, the power usage will go up inevitably, so it's better to not flat out discard such measurements in reviews.Does anyone really care?
That may take a while, if ever, if you are talking Zen3 vs Alder Lake, as I don't think there is a big performance win for Zen3 in real-life Multi-threaded apps. Take a look at this summary of common business applications:How you spread the workloads on the PC. Basically the time spend on tasks that actually "give you money" and justify the SP performance over the MT performance.
Um... what are you talking about? I said I cared about single-core performance. By definition SINGLE CORE performance is measured using one core only, which is often the fastest core. It is not measured using all of the cores simultaneously.
Although you can overclock all cores, that doesn't mean you have to do so to get high single-core performance. By design, many of the modern CPUs will run a single core up to some limit, say 5.3 GHz. Once two cores are in use then the limit might be dropped to 5.1 GHz. As more cores are used the CPU automatically lowers the clock rate of all cores as needed. I like to use a Noctua NH-D15 for CPU cooling in my builds. This has been more than sufficient, and it is quieter than most liquid AIC's.when you overclock you dont overclock one core only you overclock all of them and the more cores you have the more heat it will generate to hit Higher clock beyond the CPU specs.