Gadgetory

welcome back to harbor unboxed now yesterday we took a rather long and quite in-depth look at the thread ripper 2990 WX and 2950 x in short the 2950 x really impressed us with solar performance across the board at a super competitive price the 2990 WX though that was a bit less impressive with mixed performance and a very high asking price currently Windows 10 is mostly being blamed for the mixed results and perhaps rightfully so we'll have to look into this a bit further and keep our fingers crossed that we actually get an update and not to just in the future in the meantime there is something else we can address something that a large number of people did voice concern for not just about our testing but the testing of most other media outlets and that was that we tested just a single application at a time we had always plan to make a separate multitasking focus benchmark and I said as much in yesterday's video while editing and wrapping up that video I began looking into multitasking performance and after an additional two days of testing we've got some more results to discuss for this test the thread Ripper 2990 WX has been compared directly with the core I know in 79 ad XE and we have five batches of multitasking benchmark results to check out all of the testing was conducted using a stock out-of-the-box setting so no CPU overclocking at all and both CPUs were cooled with a 240 millimeter closed loop cooler installed on each CPU and there was some jumping up and down in the comment section not a lot but some people were claiming that the Wraiths Ripper was causing their 2990 WX to throttle and that's why our benchmarks were were mixed but just to clear all that up the wraith ripper doesn't cause the stock 2990 WX to throttle in any benchmarks that we ran keeps very cool as we showed in a review we even showed footage of it running some core heavy workloads and it was holding really respectable clock speeds 3.3 gigahertz so that's what you'd expect to see out of that CPU so not sure where all that's coming from anyway moving on both systems have been configured with 64 gigabytes of ddr4 each using 8 8 gigabyte modules and this did a limit the 2990 WX ddr4 3000 while the 79 80 XE could run much higher but we settled with a low latency ddr4 3,200 memory to try and keep things even and I suppose somewhat realistic now the following graphs might look a little odd at first especially compared to what you're probably used to seeing a promise but I'm hoping this format will be easier to follow as it gives you all the information on screen at once anyway let's get into the benchmarks I'll explain as we go ok so first up we're gonna look at how the 2990 WX in 79 ad XE compared when running blender open data and 7-zip please note the results here show how these CPUs performing each application individually so that is to say blender was executed and then completed before moving on to testing 7-zip so these are just standard benchmark results here the 2990 WX completed the BMW 27 workload 36% faster and the barbershop workload at 25% faster when compared to the core on own processor then with 7-zip we see the 2990 WX is a hundred and sixteen percent faster for decompression work but 35 percent slower for compression work now let's redo these benchmarks but with blender and 7-zip running simultaneously sevens it will be on a continuous loop while blender runs standard batch of tests so here we see a few interesting things the 2990 WX is now 58 percent faster for the BMW 27 workload and 30 percent faster for the barbershop workload however the 7-zip performance is less favorable here the 2990 WX was a hundred and sixteen percent faster previously when looking at the decompression performance and now it is three percent slower that said though whereas it was 35 percent slower previously for the compression test it's now only sixteen percent slower basically the twenty nine ninety WX was already mighty impressive in the blender rendering test and adding 7-zip to the mix makes it look more impressive but the massive advantage did have in seven zips now go on making inferior in both the compression and decompression tests the 79 80 XC certainly appears to be prioritizing 7-zip while the 29 WX is prioritizing blender or at least this is how it's working within Windows 10 next up I'm testing with handbrake and real for ham break I have an h.264 4k 60fps video that I'm converting to h.265 4k 60fps and I'm reporting the average frame rate as well as the total render time then real bench will be running it's heavy multitasking workload on a loop the 2990 WX doesn't do all that well in handbrake and here it was 31% slower than 79-80 XC it's also not great in real bench either taking 10% longer to complete a single pass so let's combine these two tests and run them simultaneously with both applications running simultaneously the 2990 WX is now 17 percent slower than the seminar aid exe in handbrake which is a reasonable improvement but of course it is still slower however while we did see a reduced margin in handbrake we see the margin increase a little for real bench the 2990 WX is now 15 percent slower so overall it wasn't really able to make up ground on the 79 80 XC in this multitasking benchmark one of the worst results for the 2990 WX seen in our day 1 coverage was found when testing with veracrypt the one gigabyte test so I'm combining that test with one of the best results for the 32 core processor which was seen when testing with Corona here the 2990 WX was 28% faster than the 79ad XC and as we've seen it always seems to do very well in rendering workloads however for veracrypt it was 47% slower so not a great result that one but anyway let's run them simultaneously now and see if that changes anything as we saw in the first series of multitasking benchmarks AMD seems to prioritize rendering workloads while Intel favors the more memory sensitive workloads the 79ad XE took almost 6 times longer than the 2990 WX to complete the corona benchmark while it only sorry 7 percent reduction in decryption performance meanwhile the 2990 WX sorry 34 percent reduction in decryption performance and a 61 percent increase in rendering time so the 32 core processor does appear more balanced here and so it should be with 14 more calls to play with still we do have mixed results for the 2990 WX it was absolutely amazing in Corona and horrible and veracrypt and that's true whether you run these applications individually or simultaneously moving on this time we have three applications handbrake cpu-z which is a synthetic benchmark and then 7-zip here's how they all perform individually on each CPU the 2990 WX is 22% slower and handbrake 75% faster and cpu-z and again we see the same mixed 7-zip performance as we discussed earlier running all three applications simultaneously does provide some interesting results whereas the 2990 WX was 22% slower and handbrake previously it's now 29% faster so that's a very impressive result that said it does take a big hit and cpu-z but even then it was still 21 percent faster than 79 80 XE and then finally even in sevens that we find positive results for the 32 chord processor it managed to maintain a good chunk of its decompression performance and the 79ad XE took a big enough hit in the compression test that the 2990 WX was able to just edged ahead so this is the first truly great result that we've seen for the 2990 WX needs multitasking benchmarks especially given that I haven't included a rendering workload to help prop the 32 core processor up last up I have another triple header using povery WinRAR and f1 2017 Popper a provided the 2990 WX with its best result in our day 1 review and as you can see it's 36% faster in the 79 80 XE in this test win rather though isn't very cool heavy but it is extremely memory intensive so I'll be keen to see if that messes with the povery performance at all the 2990 WX was 59% slower than the 79 80 XC when testing with winrar or another way of putting it the 79 80 XC was 142 percent faster finally we have f1 2017 in here the 2990 WX was 24% slower when comparing the average frame rate of 30% slower for the 1% low result so now it's time to run everything at the same time as we've seen time and time again when multitasking the 2990 w axes advantage and render applications is only extended whereas it was 36% faster than the 79 ad XE previously when testing with povery it's now 45% faster meanwhile it was 59% slower and winrar and now it's just 37% slower it's still slower of course but a greatly reduced margin however things do go home wrong for gaming performance and now f1 2017 is completely unplayable although the 79-80 XC did produce a 1% low result of just 34 FPS this is a little deceiving as the game was very playable and only once or twice per lap did I notice a big lag spike for the most part frame rates were up around 100 FPS and it provided a surprisingly good experience the 2990 WX on the other hand was a stuttering mess and would have been impossible to play thankfully how the can benchmark did take care of all our driving for me so again mixed performance from the 2990 WX in this test ok well i'm now armed with a lot more knowledge in regards to how the 2990 WX performs than I was a few days ago when I was in the late stages of wrapping up my initial 2990 WX and 2950 x a review that said the additional information hasn't really changed that much throwing more work at these CPUs just provided I suppose more mixed results whereas you'd expect the 32 core processor to really crush an 18 core processor under these conditions it just didn't steal the performance hiccups that we've seen with the 2990 WX so far I might not all be down to the pneuma configuration the increased memory latency for half the cause and the limited per core bandwidth when fully utilized it's looking more and more like another big culprit here is once again the window scheduler which to put it plainly has been known to be a bit garbage so that being the case it is certainly possible that a future Windows Update could improve things for the 2990 WX and we have already seen this in the past with the first generation rise in CPUs having said all that if you are putting down your own hard-earned cash to buy the 2990 WX and you're planning on using it with Windows 10 then I'd rather not assume anything here and just work off the information that we currently have so taking the information that we currently have at hand into full account I'd buy neither the twenty nine ninety W X or the 79 80 X II instead as I said in yesterday's review I would get the 29th 50 X the 32 core processor just seems a bit too risky at the moment the mix Windows performance in the extreme price is sketchy enough but I'm almost certain by this time next year we'll have something significantly better on offer using the seven nanometer process so from a consumer standpoint in the twenty nine ninety WX is very poorly timed in my opinion but from AMD's perspective I certainly understand why they've released such a product in 2018 anyway as always I will continue to benchmark as updates and new information come in and I'd love nothing more than be able to get right back into it with a Windows Update so fingers crossed I don't have to wait too long as for this video that is going to do it for now I hope you did enjoy the video and if you did feel free the like button for subscribe for more content and if you appreciate the work we do in our box then consider supporting us on patreon we do our monthly live stream to myself that's a lot of fun we also have a discord chat Tim and I are very active in the discord chat we're there every day answering questions and chatting with the members and yeah it's a whole lot of fun anyway thank you for watching I'm your host Steve and I'll see you again next time