Gadgetory

video is going to be very simple basically we're looking at the Ajitha updates for AMD BIOS and weather on our gigabyte x5 sony master motherboard it mattered to go from f5c to an 11 which was the bios that initially shipped on the boards we tested with f5 see spend some discussion online if you missed it where some reviewers are seeing something we already knew about but it wasn't affecting us really where they're seeing lower frequencies than advertised buddhist and we need to talk about that briefly too but the end result depends c feet of cpu and board to board and fortunately for us we were basically unaffected but we're gonna go through the test date anyway between the two different biases because you all keep hammering my inbox and comment sections about it and i am exhausted and actually ready to die at this point but we're gonna do it anyway before that this video is brought to you by us and the gamers nexus toolkit on store documents axis net our brand new toolkit just launched and contains ten custom eight drivers for video card disassembly reap hasting and tear downs the eight core tools are made of high-quality chromium vanadium alloy steel that's built for long service life and resistance to wear during use the other two tools are carbon steel hex heads that were custom ground down for capacitor clearance on video cards all the tools are easily mounted to a pegboard or stored in the GN made tool bag for easy transport learn more at the link in the description below alright so how is keeping stuff I gave the whole team off today I was also supposed to have off I don't mean to sound like I'm complaining but I do want just think of a time when you worked 168 hours in a week think of a time when you worked like a hundred and twenty to thirty of those and remember the feeling at the end of it where you you actually aren't sure if you're going to live anymore so that's where we are but you just want to set that stage so the point is for the charts today I'm not doing anything fancy I'm gonna put a PNG on the screen and I'm gonna say the numbers and then that will be the video so to catch everyone hop basically there isn't an issue on some CPUs and some other boards especially where the and when I say some CPUs I don't mean 3600 versus 3900 X I mean literally CPU to CPU there's variants within the same skew so the issue is that it's not always boosting to the advertised boost clock under the single-threaded workloads and all court workloads totally unaffected for us anyway can't speak I guess for all of the samples but from what we've seen all core a hundred percent unaffected the results for all core 100% unaffected results for a single core theoretically can vary a bit and we'll go through some of that today but it depends and so what we worst frequency data in our initial reviews a lot of people I guess I didn't blow it up and make a big story out of it so I guess no one really noticed but we were seeing on average about 25 megahertz below advertised in some of the tasks but for the most part it was okay okay enough that it would be within variance for the most part and that's what we're gonna look at today so what we did very briefly here to explain it is X 570 master for the motherboard it's by gigabyte I am extremely happy that we chose to test quite this time because from what we've heard about some of the other motherboards gigabyte has been the most Hanuman II with the speck which is good because redoing everything would suck so fortunately we're not too affected but basically there's a few different versions of BIOS there's n 11 that's what the what I am the officially validated with the CPUs and then there's fc5 I think it's called which is what came out just before release there was an F F F 5 C I think there was an f5 d and then now there's an f5 e those are gigabytes bios's MSI has had some BIOS changes as well but at the end of the day what it really comes down to aside from some memory tweaks that aren't really relevant here its overclocking stuff aside from those what it comes down to is AMD's binary code that they give to the motherboard makers and it's called Jessa and AG ESA there's version 1002 was on the or one zero zero two was on the N 11 press initial bios we did not use that BIOS for the motherboards and then there was one zero zero three a one zero zero 3a is what we used it's on fc5 f5 see sorry and that includes the some of the fixes for clock boosting so we had those and we were we were well we were supposed to have those and I'll talk about that a bit more to you and then there's another version one zero zero three a B a Jessa and that includes the initial patch and then a follow-up patch so those are the Adisa versions that's what dictates the boosting behavior and for the most part it depends on the CPU literally CPU two CPU where you'll see some that just don't do 4600 for 3900 xmi 345 785 or 4550 we can't get a clear answer from Andy on if that's normal which is incredibly frustrating but I had to call with AMD before we went live with our before we wrote the review and said hey we're seeing we're seeing this clock behavior 45 75 is kind of the max single thread we see and we're seeing a single or multi core at whatever it was don't remember at this top my head but we were seeing multi-core told them that number I said what is the all core frequency and they were like well we don't we don't have that I was like yes you do there's an all core frequency you might not call it that anymore but it still exists so the way it works it's it's like partly thermal dependent power dependent and if you don't have the thermal limitation which we don't in test environment you don't have really a power limitation there is going to be an all core and the only reason it might be dictated differently at that point would be silicon quality which is a different discussion entirely anyway we found that all core in the single core numbers ourselves thank you very much Andy and those are in the reviews for the 3600 3900 X so we did find those we published them you're good to go on that anyway I guess I think we can just talk about the numbers here let's I don't know let's just go through the numbers I'll put some p.m. G's on the screen and then we'll if I think of anything else we'll go over it and I'm going to go outside before I don't have a chance anymore so I Cinebench the first has to help demonstrate the differences and if they emerge a Cinebench we have to do real-world testing as well of course but this helps establish expectations with the r5 3,600 CPU the results aren't just close they're identical this chart only looks at the maximum at single core clock per polling interval remember that the fastest core will change from one to the next so we take the maximum value out of all the cores and plot that as it is a single threaded workload the maximum frequency is 40 200 megahertz all tests technically it's measuring 4,200 point four but basically 4200 BC LK or base clock is 100 before anyone asks about that the r5 3600 official spec as a reminder is 4,200 megahertz our r5 3600 results are 100% unaffected by this BIOS there is ZERO difference and the results will not change we test it on f5 see but an 11 would also have been fine here f5 see also has one zero zero three a a Jessa and 11 has 1 0 0 2 and the 1 0 0 3 a B can be found on the f5 II version with patches of more of the same 4 1 0 0 3 a except for some of the other affected units not even SKUs for the ro933 seeing more of a difference than the 3600 but still not much of one the single-threaded difference in Cinebench is between 0 megahertz and seventy-five megahertz max and not frequently the average delta calculated across the entire run is forty nine point six megahertz as for if this impact results the answer is that it depends but mostly no remember that not many of the tests we run these days even games will entirely load one threat and we don't do one thread Cinebench numbers either so that doesn't matter here also even most games will push out at least two core workloads these days and so these are unaffected in our testing we can look at some real data for validation we'll skip the 3600 since we've already validated it as good data as a no point in redoing all that work and so that remains entirely 100% unaffected there is zero impact on the 3600 review which is great because we said it's a good processor so that's good I'm glad that I don't have to change that opinion although it would only go up in theory based on what people have been complaining about so there is no magical 10% performance increase to our 3,600 review for those of you who wanted there to be one I'm sorry but just be happy with the fact that the 3600 is that they have good CPU and we were very confident in making a strong recommendation for it in our video so you don't need to stuff this magical plus 10% number into our charts and mislead people because I've seen a lot of that on reddit and it's incredibly frustrating just like the CPU already does well you don't have to make things up there is an issue with boosting for some processors but it's not for all of them so don't just assume every review is going to have a magical plus 10% number here all right so I think we need to talk about some other stuff be right with v-ray the 3900 X stock CPU produces the same 0.75 minute results as in our original test that's with an 11 F 5 e and F 5 C we use f5 c for testing which has a G so 1 0 3 a B and a sorry and f5 es10 3 a B and 11 is what AMD's hunt on the board but we updated to the FC version for better support elsewhere there's no difference in this test this is an all core test so everything else that's all core which is all of the production tasks we did basically I think I think yes there might be one I'm not pretty sure - all of them are unaffected they're for all core still get another one that is frequency and not a frequency dependent notice thread heavy GTA 5 1080p GTA as a reminder is more single-threaded than every other game in our benchmark for the most part so differences would emerge here for this one our original review data published a 109 point nine FPS average for the are a 939 hundred x the an 11 pre-release review bios with a GC a 1000 to hit 112 and our data range here run to run is 112 point four to one hundred thirteen point three average this is tested across five test passes standard deviation is roughly 0.4 FPS for deviation on the average the maximum difference is a two point seven percent improvement not IV 10% improvement that people online have been speculating that is again not to say it won't happen on some reviews but it does not affect our review here there's a maximum improvement of 2.7 percent and that's with the run to run variation which would drag that down depending on which number you calculate against so this is important to go back and note of course the improvement is noteworthy and happy to point it out briefly but this does not change the review quick aside here to a lot of people are now confused us what boost even means what int says 4.6 gigahertz boost they don't mean all core it has never meant that that's not what Intel mean to either that's not the problem this is an entirely different thing all core is always going to be a lower number to maintain stability and the boost spec listed on the website has never been for the last several generations the all core boost especially for risin so that's a different issue also my audio my recorders about to die but we're just gonna switch to the on camera audio I'm not doing this today all the recommendations in the review for the 3900 X remain the same the 3600 remains strongly recommended for us the 3900 X remains recommended in the applications we recommended it for I'm not redoing the review here you can go watch that if you want to know what those were and and then I guess that's that's up to GTA so far so review itself unaffected data affected maximally about 2.7 percent in GTA 5 and 0 percent in v-ray and every other all court workload look at the last two a tomb raider plot in a difference of 1.6 FPS average of max which is more or less error margin it is a repeatable result there so we'll mention it as repeatable which means that it is probably a real difference the end improvements is maybe 1% one point one percent which is within reasonable error for most testing but ours is a bit tighter than that so that is outside of our error margins just barely this does not change anything in the review once again but it does update the numbers which are marginally higher total war 2 is battle benchmark plotted a 162 FPS average originally and is now at 150 9.2 FPS average that's a 1.7 percent maximum improvement but this game has a much wider run to run variants than others note that these 0.1% lows for instance range by 9 fps here which is within the wider error margins for 0.1% lows in this game the average FPS difference is outside of error margin so it is a real difference we have an allowance of plus or minus 0.8 FPS average in this one so 1.7 percent max here anyway the point is I suppose that there can be a bit of an improvement but not always and there's not much of one in our testing that's not to say again some reviews really way more effective so an attack we saw the whatever initial data they put up and it looks like they'll have a much bigger difference than we did but that depends on the CPU and on the BIOS and we use the different motherboard and we obviously have a difference you so just to kind of close this out I've been knocking the the comment frenzy through this a bit because I am annoyed to be frank about the you need to retest everything it's ten percent better that's not true with our testing so you can stop those comments now it's it's we got it but it's not true here so the end result then is some reviewers will see a bigger difference then we did but we're not going to see a big difference in our data which means our reviews are more or less unaffected if you want to be optimistic you can maybe average the differences here and call it a maximum average improvement of two percent in games that are more thread limited so that would be the optimistic look at it but that doesn't really change a whole lot you're talking single-digit FPS changes here and to go over this again like I was saying some outlets will see a bigger difference like an attack as mentioned but I do want to be very clear here that for those people who are more effective for the reviewers who are more effective than we were I'm obviously very happy we're not that affected because we've done enough for now but for those who are more affected please don't go like brigade their channels or their review websites because this is not a fault the reviewers this is an AMD crammed a bunch of products down everyone's throat and gave him 6days launched the product on a Sunday so even getting contact from AMD was difficult and also a holiday weekend in the US so get in contact with AMD and the motherboard manufacturers was difficult so this is not a fault of the reviewers for those who do see a bigger difference don't go brigade them it is purely a matter of the BIOS revisions the GC code changes and the individual SKUs of CPUs and gigabyte via an outlet had released some information on their own internal testing we're on the same BIOS with the same skew so thirty six hundred I think it well so same skew thirty six hundred same BIOS they were seeing different single core frequencies one to the next and so how are you what are you supposed to do about that like how do you after that so anyway that's the story our results you can look at with confidence that they are fine and if you would like to be optimistic on the gaming stuff add 2% our 3700 X review I'm not going to go back and change that will be published as is because again the difference isn't big enough and it's it's not worth redoing days of work at this point because we are totally knackered so this is you know this is this is a different discussion entirely on how to treat your media partners and your motherboard partners but anyway and all of the story is our day that all core fine and then limited core workloads maximum two percent improvement except for that one was it GTA maybe 2.7 percent wherever that was 2.7 percent in GTA 5 but the max average was 2% so I think I've said everything I need to say and I would like to go outside now because I haven't done that in a while so thank you for watching I will see you all next time you can get a strata here his exes dotnet pick my shirt a mod mat or a toolkit if you'd like to support us directly and and you can also go to patreon.com/scishow and axis I'll see you all next time