Gadgetory

today marks the first of many revisits for am the risin something that we've regularly performed for new GPUs and CPUs as they arrive to market this one comes sooner than most because of the many changes which have allegedly had significant CPU performance impact already rather than speculate and run tests with all of the things applied at once like updating multiple windows or game version simultaneously we decided to individually test Windows updates game updates efi updates and memory overclocking then combine them all together for a full picture of ryan's progress over the past month before getting to that this coverage is brought to you by EVGA gtx 1080 TI icx cards the new FTW three model which will have three fans using asynchronous control for their cooling of the GPU the BRM and other onboard components you can learn more at the link in the description below or check our previous icx card reviews for more information the testing today will isolate Windows updates for performance testing we'll also be looking at EFI testing without any changes elsewhere suggest an EFI update nothing else to not take an advantage of new EFI features then we'll be looking at memory overclocking and eventually compiling everything into a complete list so this will be a cumulative set of benchmarks that means that as we go through things we're going to be stacking on the previous set of charts so Windows updates will come first EFI next and there's EFI tests contain the Windows updates and then after that the memory overclocking tests which contain the EFI updates and therefore the Windows updates so it is cumulative in that nature the windows update in particular is an interesting one because there was some talk of Windows updates potentially impacting Rison performance in a positive fashion but most of the testing we saw was posted with potentially other changes or just didn't have initial data that we could use to base off of for how much of an improvement there may have been so we've based off of our round 1 benchmark data to see how Windows has changed over the last month or so we put the r7 1700 and r7 1700 acts back to the bench the 1700 X were the least excited about the 1700 was the one that if you remember we sort of showed as the champion for the horizon 7 lineup it was clearly the best buy for a gaming audience particularly because I overclocking basically made it equal in 1800 X so it's pretty obvious why we would choose that one in terms of price now that said we're still putting the 1700 X in here because it is the most reason that we tested and for control reasons it's the easiest for looking at data comparatively with all these updates applied because our 17100 had two issues achieving higher than a twenty six sixty six megahertz memory clock without one the newest efi updates and two without overclocking the CPU itself which indicates some kind of imc limitation on our 1700 unit that we received but the 1700 X could hit 29 33 no problems so we'll be able to look at those two models for a better understanding of how memory overclocked in to a higher frequency like 34 66 megahertz or beyond will improve rise in performance will tune memory later in the video we were using EFI 5 7:04 for the first test that is the windows update and then we'll roll the performance in with the newest version which I believe is 1002 or 1 0 0 to beta that we received from ACS directly as for the Windows versions we will be testing with windows build ending in 9 7 0 from March 22nd that's the newest one at the time of setting up this test and then that is compared vs version 6 9 3 from January 10th which is the previous version we had on the test platform as a note here we do not have updated numbers for Intel or for other and these CPUs that are not included in the numbers today so that means that any changes in this Windows build that may have impacted Intel or other and these CPUs will not be factored into cumulative charts with everything on it because we just we didn't rerun them we're really focusing on rise in here so do keep that in mind these controls in place will have synthetic tests for you on the website that will be linked in the description below along with all the other charts everything is there in written form if you prefer that we're going to be jumping straight into gaming benchmarks for this video and we'll be sticking with gaming for the entire video the rest n is down there in the article starting with Metro last light reliable as its had no updates and years we see that the r7 1700 X performs effectively the same as the version that's been accompanied with a Windows Update overclock the 1700 X is within test variance for this particular benchmark tool producing less than a 1 FPS difference and averages between Windows versions the same is true for the original r7 1700 x stock test and 1700 X Windows Update test where we end up within about 1 FPS average so no change the r7 1700 shows the same average FPS for this title the January 30th Windows 10 build post no discernible deficit compared to the January 10th or March 22nd builds they all look about the same moving to GTA 5 build 970 and build 693 produced effectively equal results between the overclocked 1700 X CPUs holding true also for the non overclocked variants the r7 1700 shows a 2% difference favoring Windows Update at this point we don't yet have enough data to definitively state whether this improvement is the result of Windows updates or is it just regular variants in testing watchdogs to paints a similar picture we have a statistically insignificant difference between builds at 693 and 970 and that's for the r7 1700 X when overclocked the stock r7 1700 x with build at 970 performs at 84 FPS average where 693 performs at 82 average like with GTA 5 this is enough to start resembling some sort of improvement but the lack of statistical significance in the overclocked values gives us pause about potential for improvement here we would need more data to make a more definitive statement as to whether the 2.4 percent improvement theoretically is caused by the Windows Update but for now this is at least consistent with GTA 5 the r7 1700 also posts a similar improvement of roughly 2.9 percent in averages with build 970 again ignoring the 1700 x overclocked performance equivalent these numbers are consistent in showing a potential improvement with the Windows Update however minor it may be our later tests will impact this in theory in a bigger way battlefield 1 starts join some repeatable statistically significant differences in our 970 and 693 numbers this builds on the story we've established with gta5 and watchdogs - but gives us greater confidence that the differences are more than just chance that said quick note here dice pushed an update called they shall not pass for Battlefield 1 on March 14th along with a later pushed version 1.08 of the game although the extensive patch notes do not indicate any performance improvements as stated by dice it is possible that optimizations were made at a game level and not listed with build 970 the overclocked 1700 X performs at 137 average we're billed 693 performs at 1:30 3.7 FPS average that's another difference of 2.5% so not that relevant but frame times are where it counts because we measure frame times and convert them into our 1% and 0.1% fps numbers we can see that build 970 post 111 FPS at 1% lows where 693 again 2.5% behind in averages post a 1% low of 87.3 fps finally we're seeing significant differences it's lesson averages and more in frame times for this particular game with 0.1% low is also posting 8 gained from 72.7 to 99 FPS and again these numbers are representative of frame times the story holds true for the stock 1700 X which runs at 135 average at 102 1 % low and 88 0.1% low with build 970 against builds 693 from the original review that's an unimportant improvement in averages but a measurable and repeatable improvement in frame times here's a frame time plot for battlefield 1 on screen now while talking to the rest of the data this is consistent with our 1700 X numbers although these improvements are not visibly appreciable to the end-user they certainly show up in charts and are important in showing promise for potential performance improvements in other titles later on this chart is for ashes of the singularity but without the escalation updates remember we're trying to test one thing at a time here not throw a whole bunch of updates into the bucket and call it an improvement that's nice but not useful for figuring out the source of the improvement and ashes of the singularity we're back to what amounts to primarily a statistical insignificance or very small difference overall we're seeing an improvement in the 1700 x overclocks numbers in b9 70 over b6 93 of roughly 1.3 percent averages with almost zero difference in frame x when you consider in 1% 0.1% low converted fps values the 1700 x stock with be 693 shows again a statistically insignificant difference this time favoring at b6 93 which more or less indicates test variants between two tested values which are effectively identical in performance there is no difference to speak of here especially not when the gap is 0.6 FPS the 1700 is the same story that's the Nanak version there's no statistical significance here in differences and frame times are effectively equal now we're on to total war Warhammer unfortunately with all this talk of constraining variables that we were forced to update this title for our benchmark by steam and that means there will be changes at a game level we reached out to our contacts at Creative Assembly to confirm if this update impact CPU performance and we have heard back from the communications team but we're waiting on confirmation from the engineering team to see if updates have improved CPU performance for either vendor if we receive confirmation will update the article link in the description below with creative assemblies information or quotes as necessary we suspect the differences you're about to see are a result of potentially the game update more than the windows update but either one is critical to show so we're here anyway the 1700 xB 970 benchmark with the game update out matches the B 693 benchmark in significant ways with the overclocked 1700 ex posting a frame rate approximately 11% higher at 152 FPS average vs. 137 FPS average same times don't show a huge improvement in 1% lows and the 1700 X stock goes from 128 FPS average and 94 fps 1% lows to 145 average and one 10 FPS 1% lows another significant improvement we replicated this on the r7 1700 which moves from 120 FPS average to 138 FPS average with lows also improved and here's an interesting thing total war Warhammer was also the game that showed the biggest difference with he disabled versus enabled so the next thing to look at would be how do these numbers align with the original numbers with SMT off and how do they align with SMT on and the windows end game updates this chart contains SMT 0 values for the r7 1700 here we show that build 970 more or less makes up the difference in SMT toggling indicating that one of the two applied updates could have removed the SMT hamstrung at least part of it that we previously observed now of course the question is still whether SMT toggling benefits in this title even with these updates and the answer is yes with the performance ceiling still about the same as previously at around 140 FPS average er 7 1700 is still better with SMT off but the differences are now minimized the performance ceiling remains the same in total war as previously but we're closer to it without requiring disablement of SMT we're moving on now to the efi updates on the acs crosshair 6 platform we previously did test some efi version updates with our just research Rison article that's what it was called and there was also a video on it we showed really no difference that was significant in any way on the version we tested but this one's a lot more advanced than those so we'll be going through it we are on these same windows update build as the previous tests laid out this will be a quick section with largely uninteresting results and we're moving from at version 5 7:04 to version 1 0 0 2 on the ACS crosshair 6 after which we'll do the memory overclocking with the r7 1700 Metro last light shows no benefit from strictly performing an EFI update on the crosshair 6 we need to actually use updated features as agiza updates were not enough to impact performance here in GTA 5 moving on already we're seeing 0 noteworthy change from build 970 with five seven zero four to build 970 with one zero zero two moving on swiftly to watchdogs 2 we're seeing no significant change from B 970 with five seven zero four to be 970 with one zero zero two we are with invariance at this point that'll kill one same story the improvement is clear from Windows Update but there's no clear improvement from throwing an EFI update on there as well using the new features sure but not standalone the story continues for synthetic tests and for other Gamze benchmarked it's really not interesting at all so we're going to move on but basically what we have so far is a potential improvement from either Windows updates or game updates in battlefield one and total war Warhammer the other games did not have updates as stated Windows updates and the forced game updates are most important in those two titles the f1 in total war Warhammer the others are largely insignificant differences for the most part and it appears that the games right now at least these to be f1 and Total War are around where performance was when SMT was disabled in our original review so that's looking better though you can still disable SMT with these new updates and get better performance in stock the difference between those numbers is not as great in these two police in the Total War title which had the biggest difference overall the ceiling itself though doesn't look higher with the 29:33 mega it's memory or 2666 on the 1700 so let's move on to memory overclocking to see if there are real gains so a few notes here if you're not familiar with memory overclocking on Rison or Intel memory overclocking and timing tuning right now is not necessarily trivial with Rison with the newest efi updates it is certainly easier but the process isn't as simple as if you're used to an Intel platform flipping an accent piece which going to a frequency dialer and increasing it and hoping that it will probably work as long as you don't go too crazy it's not that simple with Rison right now what we had to do was memory training so we step through memory frequency improvements incrementally one for the next established stability and then did the next set of updates and increments after sorting through six kits of memory and spending about an entire day on this task we found that our Trident z30 200 megahertz CL 14 kit seems to work the best right now and that is because it has Samsung B dies where the other kits like the HyperX kits which largely use hynek's thighs generally have issues with overclocking or even achieving their stock capable frequencies as advertised on the sticks with Rison and riding 7 and the motherboards that we have including gigabyte and Asus and this is something that we've talked about previously so B dyes are where to go for Samsung from Samsung for these memory kits if you want to achieve a higher clock you need to start there g.skill has a lot of them that are known to be bend as di specifically we've also since requested an even higher bin of memory from g.skill so should have further updates on this front for you soon but for today we got stuck around 34 66 megahertz and we're awaiting additional shipment from G scale to push further so thanks to EFI updates direct support from build Zoid and acs and support from g skill we were able to walk through the six or so hours of memory training required and tune in to establish a stable overclock of 34 66 megahertz with CL 14 timings we were able to hit thirty six hundred megahertz and tried pretty damn hard to stay there but it just wasn't stable for actual load usage with the hardware that we have available apparently and we'll have to try other CPUs and other memory kids going forward a lot of tuning is figuring out whether your are seven chips respond better to reference clock tuning or memory multiplier tuning and then going from there figuring out if you want to operate from a base of 24 26.6 6 or 32 X for the multiplier in the memory this is mostly a guess and check game which takes the majority of the time required after you've established those baselines for reference clock and memory multipliers you should be able to get a stable OC that will run without any issues we're using the 1700 X for this since all the most recent tests were conducted on that chip and because I had a 29 33 megahertz base for all of those tests making it very easy to do the comparison the 1700 as stated was a lower frequency for the non overclocked tests that we did and hire for the overclocked ones making it less linear to compare will have these synthetic and rendering benchmark results in the article below as stated previously and we'll just go through gaming now starting again with Metro last light the 34 66 megahertz 1700 X with stock clock sees a performance bump from 123 to 125 FPS average with lows effectively identical around 81% and around 50 10.1% that's an improvement of about 1.6 percent by going from 29 33 to 34 66 megahertz with the Windows updates and the EFI update in along with the memory overclock so we've seen about a 1.6 percent gain here looking next to court overclock and we see the 1700 X at 3.9 gigahertz with 34 66 megahertz memory is placing around 129 FPS average or roughly 1.9 percent ahead of the 1700 x @ 3.9 gigahertz with 29 33 my guards memory the change being only in the memory compared to the original results which had be 693 and 29 33 on the 3 point 9 gigahertz overclock we have moved up about 2.4 percent from 120 5.7 FPS average to 120 8.7 FPS average you might as well add these into the original tables as well for comparison note that the Intel series and the 1800 X haven't been retested with the new Windows or game updates so anything which impacts Rison could theoretically also impact Intel or the 800 X we haven't yet determined to this on this table we're seeing our best effort 1,700 exit line item with a core 3.9 gigahertz OC and memory 34 66 megahertz OC finding around where the original reviews are 7 1700 4 gigahertz OC with SMT disabled landed we are below the i7 4790k stock CPU which sits at 1:30 6.7 FPS average with the 1% at worst frame times converting to an FPS of 101 and 0.1% worst 294 FPS the 7700 K modern flagship is at 146 FPS average with 104 and 95 1/4 70.1% loads representing the frame time the 1700 X with its memory OC and core over clocks run at 88 FPS average of 1% low and 56 FPS average at 0.1% close disabling SMT as illustrated by previous tests improves frame times in this particular title back to GTA 5 to the next one here we're seeing our 1700 X at 3.9 gigahertz with 3466 mega its memory performing at around 135 FPS average with an improvement in frame times over base the 1700 X at 3.9 gigahertz with 29 33 megahertz memory is that around 128 FPS average is showing an improvement of about 5.7 percent by overclocking the memory 234 66 megahertz if we kill the core OC removing the core OC and just run the memory OC we're at roughly the same place 120 an FPS average with lows around 80 versus a stock 1700 X with 29 33 megahertz memory the 3.9 gigahertz version with 34 66 megahertz memory improves performance by about 11 percent again that's by overclocking memory and core to compare it to a simplified stack that puts the 135 FPS 1700 X above the 1,800 X when it was overclocked with builds at 693 or the 1800 X's best-case in other words it is also below the iPhone 4790k and averages and frame times which leads by about 4.4 percent average the overclocked i7 7700 K is next to be compared to the overclocked 1700 X the 7700 K is similar in price to the 1700 and for pure gaming builds it should be alongside the 1704 head-to-head comparisons keep in mind also that the 1700 can support the same overclock and memory overclock as our compared 1700 X which means performance is identical at the same clock the 7700 K overclocked is at 1:51 FPS average with 101 and 93 fps lows at 7700 K still therefore holds a performance lead of about eleven point nine percent which is down from its previous performance lead of 18.9% so there is improvement here that said we have not yet retested the 7700 K with build 970 of Windows 10 or with memory overclocking of its own and the 7700 K here is at a 32 hundred megahertz CL 16 memory clock and timings and watchdogs to the stock 1700 X with 34 66 megahertz memory places at 100 FPS average with lows in the 70s and 60s vs. the 1700 X stock with 29 33 megahertz memory this is a significant improvement of 19 percent so far all of our improvements in the same comparison have been in the single digit percentages making watchdogs to a bit of an interesting title we also know that watchdog 2 is particularly thread intensive and previously published a video and article studying the CPU impact from different settings in the game and so there could be some sort of cross CCX traffic that benefits from the faster memory in this case well you can have the 3.9 gigahertz 1700 X at 29 33 we're faced with an average FPS of about 87 which previously showed no note where the improvement over the stock 1700 X with 29 33 megahertz memory memory is clearly choking rising in this specific title and looking against Intel the 7700 K overclocked lands at 114 FPS average at 91% low frame times and 76 0.1% blows the 1700 X with core memory OC lands at 107 average 75 1% low and 60fps 0.1% lows again we'd have to overclock Intel's memory for final validation of this title but this gives an IDs where aizen's memory overclocking performance lands back to battlefield one it appears that the most important item is the B 970 Windows Update and or update the battlefield one since previous testing after which point overclocking becomes less significant the 1702 x with its 3.9 gigahertz OC and 3466 memory posted no significant statistical difference versus the 7200 X at 3.9 gigahertz with 29 33 megahertz memory the difference here is effectively zero percent battlefield 1 like watchdogs 2 has its own unique performance needs looking at the stock 1700 x vs the stock 1700 x @ 3466 my cards memory so the memory is o seed we also see an effective equal average frame rate I'll be it with improvements to the 0.1% level metrics from 86 to 90 5 fps back on the bigger charts with the other tests and CPUs this lambs the new 1700 X metrics around the I 560 600 K at 137 FPS average 89 at 1% low and 77 0.1% low is where the 17 100x has a lead in frame x over the i-5 but is around the same in averages versus the nearest i7 the 1700 X is again under the i7 4790k which is at 140 FPS average for the next chart again we haven't updated Ashes yet because that's not the point of this test ashes appears to care about memory overclocked more than core overclocked in this test configuration the 1700 X at 3.9 gigahertz 34 66 megahertz memory places at about 41 FPS average just ahead of the 39 FPS average of the stock 1700 X with 34 66 megahertz memory the latter of the two with its 39 FPS average post an improvement over the stock 1700 with 2933 records memory of 18.8% versus our best of be 693 metrics this is another title that seems to make use of the gains again we haven't rerun Intel yet or overclock its memory so mileage may vary until we get those results in but comparatively these 1700 X is now price competitive with the 7700 K in this title as the 7700 K is running 43 FPS average with its core overclocked or 42 without we're curious to see if a memory OC benefits kb lake in this particular game if we were to spend the same amount of time working on that but for now the 1700 X with all of our overclocking effort is starting to claw back some ground finally we have total war Warhammer again keep in mind that this game it just receives a patch and we haven't yet rerun the other tests including Intel or FX benchmarks for the r7 1700 X overclocked in memory 234 66 megahertz with a stock 1700 X provides a performance boost to 160 FPS average and 1 24 fps 1% lows this is up from the stock 1700 actual 29 33 megahertz memory at 145 average for performance improvement of roughly nine point nine percent versus D 2.9 gigahertz 17 Ardex 229 33 megahertz of memory and B 970 Windows we see an improvement of about five point one percent the core OC and memory OC puts is up to 165 FPS average for a gain of about eight point eight percent over the overclocked 1700 X with 29 33 megahertz memory and now test it against the other items on the chart the 7700 K overclock that lands at 187 FPS average with the stock 7700 K at 186 FPS average it is possible that we're bumping up against other limits here given the proximity of performance between these two line items the overclocked I 570 600 K runs at 177 FPS average with poorer frame times compared to the 7700 K more similar to the stock 1700 X and it's best white with all the over clocks the 1700 axle ends and the ahead of the 6600 ki 5 and behind at the 7600 ki 5 and averages the lows are ahead at 1:12 there's a lot more to do as always but we'll leave that here for now there's ton of data to sift through and for today that gives us a baseline at least for improvements from more effort in overclock and then might be normal for these types of benchmarks my thanks to our team for helping Patrick spent the better part of a day really establishing a stable memory in core overclock on the 1,700 acts with all these changes and Andrew will undoubtedly be spending a long time editing this video with all the charts after the overclocking effort note here that we did have to keep the crack in X 62 on there that cooler was needed to ensure that these 1700 X did not encounter thermal throttles when testing because in our original review we showed that it is possible the thermal throttle that CPU if you're using a worse cooler and you're blasting voltages to try and sustain higher clocks or something like that or just generally if there's more traffic on the cores between T C X's or otherwise just in general more traffic you might have thermal problems so we threw a cooler on there to ensure that was not a concern we do that with all these tests unless we're testing thermals and there's more thermal testing coming soon though so I'll leave that for that future content for now we've got you the updated FPS numbers we have you the data to sift through it's been provided look forward to future testing there's synthetics in the article linked below thank you for watching as always you can go to patreon.com/crashcourse as' if you'd like to help us out directly and purchase us something like a bot that automatically reads through these scripts and edits the videos I'm sure that's possible right with the patreon money we have otherwise subscribe for more thanks for watching I'll see you all next time you