Gadgetory

if you haven't seen our rise in content since the initial are seven reviews note that we ran a risin revisit series which found game updates had improved gaming performance in some titles we'll be including that data in our new tests alongside Intel retests for the same game also in case you missed it our VR benchmark for the r7 versus Ison prize our first in-depth virtual reality test of the rift and vibe will be performing similar tests for the r5 CPUs this week for now though we're reviewing the r5 1600 X and 1500 X CPUs in the risin family it priced at 250 and $190 respectively before getting to that this coverage is brought to you by the 1080 FC which has a new MSRP that is it lowered with the launch of the 1080 TI series at 1080 and 10 tick the SC cards come with for honor or Ghost Recon wildlands which you can choose that check out learn more at the link in the description below just to recap everyone on the horizon at 7:00 revisit content we tested a few things there standalone EFI updates which yielded no performance benefit without taking advantage of the higher memory frequency permitted by those EFI updates we tested the higher memory frequencies up to 30 466 at the time that we can now hit 3,600 on some platforms more on that later and then we also tested Windows updates standalone which yielded no performance benefit going from be 693 to be 970 didn't magically fix things in fact it was the game updates which happened to coincide with the windows update that actually improves performance and for total war Warhammer that was noteworthy for battlefield 1 there was a performance uplift these performance uplifts also applied to Intel which we have now partially retested well note in the charts which ones haven't been retested yet so we've got that we retested Rison 7 for this benchmark of our fives just as a note so these are not recycled numbers from the original review and we have the EFI update on the crosshair 6 along with 3200 megahertz memory at 4 risin 5 for the horizon 7 retest in gaming but for these synthetics and blender we're still on 29 33 megahertz for the RS 7s we're on 3200 for the our five and then Intel has remained the frequency is defined in the article link in the description below which contains all the test notes and some additional tests in the synthetic and gaming categories we've got the CPU specs on the screen now at least the basics these 1600 X is a 6 chord 12 thread part with a 3.6 gigahertz at base frequency and 95 watt TDP at $250 the r5 1500 X our other review part is a four core eight thread unit at 2.5 gigahertz base we don't yet have the r5 1400 on this chart or r5 1600 cpus and because of these new our 5 CPS are still using 2 CC X's just with some of the core is disabled that means memory frequency will still matter here when the Infinity fabric as the AMD calls it lags behind in cross CCX transactions so memory speed will still matter quite a lot for our 5 CPUs it's not just in our seventh in because we're not looking at a single CCX for these units it's two of them for the most part with cores disabled that noted the methodology as always we linked in the description below in the article we have updated that with our rise and revisit info the VR stuff is coming shortly a bit later from this because it needs to be a completely standalone video it's very in-depth stuff and then a final note here we are adding error bars to some of the charts here it was added for our VR tests people liked them a lot so we're going to go ahead and continue a pilot run of adding the error bars they are calculated over multiple passes so we take all of the data we've collected over the past year or two really and calculate the error calculate margin of error calculate the standard deviation all that stuff look at it statistically and determine what is the variance test to test either from a game from the OS or whatever other variables are involved in the system thermal testing on the r5 1600 X shows that using our crack in X 62 cooler we're not in danger of encountering a clock throttle under torture workloads at any point during this test this was not true for some of the previous rise in CPUs like the 1800 X and our 7 1700 X units which had high V core out of the box and a higher temperature as a result the 70 100 M 1600 X however are able to keep a flat clock when under thermal torture test including the 1500 X by the way the sporadic lines at the front of the chart illustrate X F hour prior to test execution if you're curious we're still working on running separate thermal tests for Rison and we'll be publishing a complete standalone piece that completely defines thermals we've worked with AMD on this as well as others and that will hopefully put forward a complete picture of what Rison looks like for temperatures because it's been very confusing thus far as for voltages and power the 1600 x drives about the same power from the wall as all the other rising ships expectedly as of the 1500 x under a blender multi-threaded workload the 1600 x stock cpu clock set 2.7 gigahertz across all cores when working concurrently on all those threads and that runs the core voltage of that 1.17 to 1.2 power draw is 145 watts constant with this workload compared to the 123 watts constant on the r5 1500 X same workload with its 2.6 gigahertz speed and 1.2 to 1.2 for V core Auto controlled compared to levy 1800 X to about 185 watts in the same test with a 1700 X drawing 168 and 1700 stock CPU trying to 133 pre overclock now if you overclocked it it'd be a bit more and it would also equal the performance of the 1700 X and 1800 X CPUs let's move on the competitive benchmarks starting with blender premier and then move on to gaming we're running our blender tests using release candidate ready settings with 400 samples per pixel and 16 by 16 tiles this is the optimal use case for blender with CPU rendering and one with which we're familiar from our own animations that we rendered in-house for the channel as is known a blender cares the most about threads and so there should be significant differences between the are 512 thread and 8 thread chips here despite likely minimal differences a clock for clock in game we're seeing the r5 1500 X stock renders the 4k scene and fifty five and a half minutes faster than the overclocked I 570 600 K by about seven minutes that's an 11 percent reduction in total render time the next closest CPU is the i7 2604 points having gigahertz which completes the render in fifty four and a half minutes really showing Sandy Bridge is proud us here in overclocking the 1500 X 2 4 gigahertz leapfrogs the 2600 k OC landing at 50 points seven minutes at which point we're in Devil's Canyon i7 territory these 1600 x stock leverages its extra 4 threads here completing the render in thirty six point eight minutes compared to the 1500 x stock that's a render time reduction of 33.6% note here that our r7 CTS haven't yet been rerun at 3200 megahertz though you can see the memory benefit between the 29:33 1700 x OC and the 1700 x OC 134 66 megahertz that should allow some extrapolation it's not as important and this type of test as in some of the others price the performance intel's I 5 CPUs don't stand up here the i5 7600 K overclocked to 4.7 gigahertz finishes its render in 62 and a half minute the r5 1600 X overclocked to 4 gigahertz finishes in 32.7 minutes that's a render time reduction of 46 percent for roughly the same price if you actually do plan to do some CP bound rendering the r5 and our 7 chips remain as we have said from the get-go achieve consideration at the price let's look to Adobe Premiere video rendering benchmarks now we've included cuda accelerated benchmarks as well just to provide perspective for users who prefer cuda accelerated rendering to software rendering we are rendering an EVGA icx review from the 1080 ftw2 that we uploaded on the channel so that includes some features like color correction some warp stabilization things like that in there we're running with a GPU for the 1600 X 5900 k8 hundred and seventy six hundred K all of which land within a few minutes of each other when using the GPU to render so the extra threads on the CPU don't help a whole lot in fact generally having a higher frequency helps keep the CPU up with the GPU but overall we are no more than a few minutes apart even from the 7600 K to the 6900 K looking instead the CPU rendering the r5 1500 X finishes the render in 114 to 126 minutes depending on the overclock which is 22.7% behind the 12 thread r5 1600 X at 84 minutes compared to an r7 1700 overclocked which we still think is the best buy of the r7 family the overclocked r5 at 1600 is about 28% slower if you want to see more benchmarks we have pov-ray Cinebench time spy and fire strike on the website link for the description below and those contain additional synthetic tests battlefield 1 is one of the games where we saw performance to uplift following the recent patch 1.08 which improves performance on both Verizon and Intel CPUs by the way our eyes and revisit showed r7 uplift but did not revisit the Intel CPUs at that time for this review we've read benchmarks the i7 7700 case ah can overclock somebody sick kind of case tockman overclock and the i5 2500 cpu other Intel CPUs have not been retested in this title and will so place a few percentage points lower on the charts than if they were retested this is just a matter of time limitations at this point all AMD chips and the most relevant Intel chips have been retested on the 1.08 build of them in the FX series the AMD r5 1500 X places on the chart at 126 FPS average with low is at 81% low and 63 0.1% lows surrounding the 1500 X the Intel i5 4690k runs at 132 FPS average with 83 fps 1% low and 72 0.1% lows this plants at the 4690k about 4.3 percent ahead of the 1500 X when both are stock the 1500 X outperforms the i5 35 70k Ivy Bridge CPU which runs around 125 FPS average and the i5 2500 k1 overclocked around 124 FPS average the i5 7500 CPUs and these closer price competition from this generation for the 1500 X and that operates at around 138 FPS average 87 and 77 fps lows that lands the i5 7500 about 9% ahead of the 1500 x with a $10 increase in available price the i3 72 50k further reinforces our initial reviews point that although the unlocked I 3 CPU is really cool from the perspective of being an unlocked i3 CPU $180 price point doesn't make much sense and that holds true when looking at both these 7500 and the r5 1500 X the 7500 or 1500 X both make way better suns in this price gap looking at the r5 1600 X now the CPU plays at 135 FPS average with lows at 98 and 86 just behind the r7 1700 stock CPU this proves that games don't much care for the extra cores in fact you'd most often be better off losing some threads and just overclock in a cheaper chip if gaming is the desire overclocking the 1600 X gets the CPU up to 136 FPS average at 4.0 8 gigahertz I was clocked quite high which was right around where the 1700 X up 2.9 gigahertz 3466 my cards memory landed given a $150 price disparity between the two if gaming is your thing the 1600 X effectively invalidates the 1700 X and 1800 X in this title we'll have to look at others for a better understanding though now the update didn't only benefit AMD Intel vy5 7600 K also saw uplift now placing at 144 FPS stock with 146 FPS overclocked battlefield 1 doesn't seem to care too much with our settings that lands the i5 7600 K stock around 7% ahead of the 1600 X stock with lows effectively equal the overclocked 7600 K is also roughly 7% ahead of the overclocked 1600 X moving the total war Warhammer this is another game where game updates not Windows updates as some people seem to think originally provided improvement for AMD CPUs upon running retest we also saw performance uplift on loads Intel CPUs and we've designated the CTS that re tested with an asterisk CP is lacking the asterisk word potentially see some performance uplift if retested by 770 700 K is running into an external constraint as indicated by the lack of the overclock scaling this is a limitation elsewhere in the bench and you can see that by looking at the positive scaling in between the 7600 K and its overclocked Experion this means that a rough 190 FPS is our performance ceiling in this bench course with our config our 5 1500 X stock averages 142 FPS with lows at 85 and 73 this place is the r5 1500 X ahead of the overclocked i7 2604 point 7 gigahertz running an average of 130 F guess with lows at 95 and 89 and markedly ahead of the i3 72 50k in frame time performance by comparison the stock I 5 7500 a locked CPU performs at around 150 FPS average with lows at 94 and 85 thanks to Lee's update before the update the CPU struggled with low-end phantom performance the 7500 leads of the 1500 X an average frame rate by roughly 11% the 7500 also leads an average at milliseconds between frames but as this frame time chart shows the difference is imperceptible to the end-user we're talking 85 versus 73 for the one percent slowest frames here not a huge difference at that point overclocking the 1500 X gets the CPU to 151 FPS average closing ground on the 7500 the 1500 X outperforms of stock 1,700 X and 1700 at once overclocked and is effectively equivalent in average frame rate to the 1600 X despite a slight frame time disparity the r5 1600 X stock performs equivalent to the 1800 X stock aiding the overclocked 1700 and further challenging the existence of the 1800 X for our audience to 1600 X overclocked to 4.08 gigahertz outperforms the stock r7 CPUs all of them and is effectively equal in averages to the i5 7500 though outperforms the 7500 significantly in low values by way of comparison the overclocked i5 7600 K is about 15.6 percent ahead of the r5 1600 X overclocked and averages with low is more or less equal the non overclocked I 570 600 K does worse in lows at than the r5 1600 X stock but the average is maintained at 174 versus 151 fps favoring Intel watchdogs 2 has proven itself to be more reliance on thread counts than other games on the bench to some extent but does still care about clock speeds it's just that the i7 CPU is post a distinct advantage over the i5 CPUs and so we should expect to see the same scaling or similar with Rison the r5 1500 X stock Seaview performs effectively equally to the i5 4690k Devil's Canyon CPU and outperforms the i3 7350 KF 5 gigahertz this again further proves that the i3 73 50k at its price is really not a good deal in just the same way that the more affordable Rison chips are effectively invalidating the 1800 x in gaming considering overclocks the 1800 X $500 doesn't make sense for gaming when you can buy an r5 overclock it and get the same or very similar performance versus the price Comparative i5 7500 we're looking at a frame rate of 80 FPS average on the 7500 9k CPU with the overclocked 1500 X at 3,200 megahertz Ram landing at 75 FPS average considering the price difference and production advantage assuming your on mixed workloads this closing distance places the r5 chip in the best price to performance brackets of AMD's current offerings moving to the r5 1600 X this CPU lands at 84 FPS average or about 15% ahead of the stock 1500 X overclocked in the 1600 x gets it beyond 1800 x stock levels just below the overclocked I 720 609 the i5 7600 K 4.7 gigahertz tests a 93 FPS average the 7600 K overclocked holds the lead of about 5% over the overclocked 1500 X stock to stock we're looking at a performance lead of effectively zero percent the 72nd K and 1600 X are equal when they are both stock and moving now to Metro last light useful as it hasn't been updated in years which is nice we see an interesting scenario that resembles the total war Warhammer game scenario is just a few weeks ago by 5 CPS have trouble in frame time consistency with this game they always have and we've shown that for years at this point and disabling hyper threading on i7 CPUs reinforces that though they do okay and averages r5 1500 Xbox TV performs are 113 FPS average with loads at 78 and 63 this is similar to the stock i7 2670 Bridge CPU in both frame times and averages overclocking the 1500 X 2 4 gigahertz lands its performance right around where the stock r7 1706 by 5 7500 CPU operates an average of 114 FPS with sporadic enough frame times to result in 30fps 0.1% low values the r5 1600 X meanwhile is effectively equal to an R 7 1700 ax and performance showing that for games where thread count isn't as relevant which is most of them there's not much tangible gain from an extra 4 threads on Rison the i5 7600 k overclocked and that's an FPS that is higher but reflects the same sporadic frame time performance as the other chips discuss Intel Core i7 chips meanwhile maintain a healthy lead in Metro last light including the i7 2604 point 7 gigahertz let's just take them out and point that chip out again we did it revisit recently and it's holding on very well the r5 chips look good in this benchmark when compared to the price equivalent I 5 chips from Intel this first Ash's benchmark shows the old version of the game the next chart will show the escalation updates which has fewer data points but it has been tested to some extent ashes of the singularity with high settings plants the r5 1500 x stock at about 26 point 6 FPS average 19 1% was in 1801 percent blows this is effectively identical to the i7 2670 performance overclocking the r5 1500 x2 4 gigahertz doesn't really gain much ground moving 228 FPS average and we proved in our previous rise and revisit piece that Ash's cares a whole lot more about memory overclocked than core overclocked if you wanted a high mark in this game if you better off investing your time in memory than core we've left our 34 66 megahertz r7 1700 X overclocks on this chart to illustrate that that you should check the rise and revisit content for a better understanding compared to modern CPUs the Intel i5 7500 performs at 26.4 FPS average with lows from 18 to 20 this is again effectively identical to the 1500 X stock CPU the r5 1600 X is a little more interesting performing within 1.4 FPS of the i5 7600 k stock overclocking both be 7600 K and 1600 X maintains a similar gap the i5 7600 K is now at around 32 point 6 FPS average with the r5 1600 X around 33 point 4 FPS average more interestingly the r5 1600 X overclocked performs it roughly equally to an r7 18 or deck stock even with the same memory speed again this further emphasizes our point that the r7 1800 X is a bad buy for gaming at the price if you are building a pure gaming machine as stated previously you're best off with either Intel the r7 1700 or now one of the r5 chips we'll talk about this more in the conclusion though the 1700 and 1800 X as shown in this game and so the others don't offer very good value for game performance particularly when it considering overclocking the r5 we haven't retested everything in Ash's escalation at this point it's more AMD heavy than Intel heavy right now since we just started running these tests with the r5 CPUs the 1500 X is now around 29 FPS average with lows in the 19 to 21 range this is effectively equal to the i5 7500 again the r7 cpus were iran here but we haven't put the 1600 x3 just yet there's no argument here that for the price Rison is the best competitor for render workloads on the cpu so GPU accelerated rendering does still serve as somewhat of an equalizer when looking at even i5 r 5 sr 7s and i7 as we show with the cuda accelerated premier benchmarks so if you use compatible benchmarks with GT rendering keep that in mind if CPU rendering is you're thin then Rison 5 is well ahead of same price I 5 CPUs from Intel for gaming Intel ties in some games like watchdogs 2 or rather an D ties Intel as it is the Challenger here and Intel is about seven to 15 percent ahead in other games averaging about seven to ten anything more than that is less common but does happen and Andy in this way has closed the gap significantly compared to its r7 vs. i7 launch that we already looked at so to that end it looks like r5 offers an even stronger argument for users who legitimately plan to do some sort of mixed production workloads and when I say that what we're talking about is looking at things like someone who might do some content creation for YouTube maybe twitch or whatever animation in addition to doing some gaming if you're not doing any animation any production on your pure gaming we have some notes for you later in this conclusion now with regard to frame times and these are five at CPU is that worst roughly equal to Intel i5 CPU and at best a good deal ahead in games like Metro last light or pre update total war Warhammer though the update has shown that yes game updates can significantly improve frames on consistency on CPUs even when it's two different vendors and Intel in this case extra threads help over the i5 CPU is but don't really do a whole lot for you over the 70 PS so the extra four or so threads in our seven really not a whole lot of gain versus the r5 if you're looking at similar clock scenarios or overclocking over i-5 it is relevant just like the i7 over i-5 is relevant it seems that 18 sweep out where it matters and also seems to be about where it stops in terms of relevance for the extra or high thread count CPUs if you're purely gaming and not looking in three hundred plus dollar territory it's looking like are five CPUs are close enough to i-5 to really justify a purchase if only because the frame times are either equal to competitive i-5 chips in price or at least slightly ahead if not a good deal ahead depending on the game now the citizens are always visible as we've mentioned a few times now if you're looking at 70 versus 85 when you're talking up the slowest 1% of frames that's not necessarily visible or perceptible to the user but it's a significant enough difference that it can matter ensures that it does matter in some game if you are looking at something like an r5 1600 X overclocking it would more or less invalidate a lot of the r7 purchases for pure gaming that does not hold true if you're doing something like blender alongside gaming but if you're pure gaming the r7 CPUs even the 1700 which we've liked do not really make a whole lot of sense when you can get a competitive or comparable r56 core 12 thread or something like that and overclock it those games won't really leverage the extra threads for the most part though render tasks will so it just depends on what you're doing the 1600 X / the 1700 would save you some money and we would strongly recommend taking that money and putting it toward better Ram because as we've shown in our rise and revisit getting something like 34 60 or six megahertz memory or 3600 and I don't just mean the kits that you can buy you have to actually put in a lot of work to get there and make sure it's Samsung B die there are a lot of requirements here so it's not just biotin it works but if you can get it working it makes more sense to put the money towards the memory because faster memory with your 70 or $80 savings from an r5 over an hour seven CPU would put you further ahead than just getting an r7 CPU with lower price for the memory now that said you have to be careful not to venture into AI 777 hunt a territory with your price as soon as you're putting so much money into the r5 platform memory and all of that that you're rivaling i7 territory and you're only gaming it's looking like the i7 7700 K is still the best gaming chip out of the ones we've tested and that is done with thirty-two hundred megahertz memory quite easily so yes I 5cq still provide a decent experience for gaming especially but even for gaming it's starting to look like you're either buying a 7700 K now because it is significantly ahead of the r5 CPUs and it's pretty well ahead of the r7 CPU or if you're not buying the 7700 kctu you're buying an r5 CPU we don't see much argument for our 7s in gaming at this point although there is an argument in some cases and we also see a fading arguments for i5 CPU s the current juggernauts are interestingly the i7 7700 K at the top end and the r5 1600 ex particularly with an overclock which is quite easy to do if you're not pushing memory very hard if you're already settling for an i-5 from an i7 it's not much of a jump to go to an r5 and ensure better frame times going forward and the average deficit is not that large 7 to 10 percent on average but you're still in good range for playable frame rates and the frame times hold on quite well by 5 is still good CPU don't get us wrong it's just not compelling enough it's not as strong as the i7 is against the r7 i7 poses a very strong defense to the r7 in gaming going beyond 8 threads doesn't do a whole lot for your gaming experience but as we've shown numerous times and i-5 reviews going beyond 4 threads does help in consistent frame time delivery it's not required you can still have a good experience without 8 threads in most games but this is actually becoming a noticeable thing in some titles that we test 16 threads don't matter as much and won't anytime soon but again it does and if you buy an r5 overclocked it by decent memory it'll be competitive with Intel just ganna be wary of entering into seven 700k territory because at that point you'd be better off buying the i7 one final reminder here it's not just Coors doing this if Coors were the only thing that mattered the FX series would have been a whole lot more competitive and less of a blunder overall looking at Coors is kind of a difficult thing to do if you're just looking at the number a core does not necessarily equal a core when you're looking across architecture that is troopers n this trooper bulldozer that's true for Intel so when we're looking at the core count people always talk about having a lot of threads available having eight threads sixteen thread that's not all it is it's having the threads available with an architecture that supports them for the task you are performing Verizon it's looking like these extra threads eight and onward are a whole lot more valuable than they were on FX and that's because of the architectural differences it's because of extra floating-point focus it's because of better AVX improvements as we talked about in the r7 eighteen hundred extra view in the article so that's just kind of a thing it's not just a core argument frequency matters two threads matter but all of it comes down to the architecture and how it works together and then we have some weird things like the Infinity fabric and memory permitting some faster transactions when looking across you see X so every architecture has its own quirks that's all for this one though hopefully that really sums it up for everyone check back soon for our continue to be our benchmarking coverage that will be on the channel shortly thermal testing for the horizon cpus gtx 980ti a/b partner reviews kutenai Messiaen gigabytes and make sure you subscribe for all of that if you like our type of testing as always you can go to patreon.com/scishow and resume excellence helps out directly dan razaaq's has done that for the full article store doc gamers Nexus tonight if you're a fan and you want a shirt like this one that's all thanks for watching I'll see you all next time you