Comments Locked

75 Comments

Back to Article

  • The Hardcard - Thursday, July 11, 2019 - link

    On the SPEC page, are those the first scores, or the updated ones?
  • Ryan Smith - Thursday, July 11, 2019 - link

    Those are the updated scores.
  • PyroHoltz - Thursday, July 11, 2019 - link

    Like Intel, does AMD produce any reference boards, that could go along with a CPU launch, to mitigate this type of SNAFU.
  • lioncat55 - Thursday, July 11, 2019 - link

    Does intel release any of the reference motherboards? I am guessing AMD makes mobos for early tests internally.
  • jeremyshaw - Thursday, July 11, 2019 - link

    I don't believe Intel produces reference boards for consumer use, anymore. They stopped before Skylake, iirc.
  • ballsystemlord - Thursday, July 11, 2019 - link

    Actually, the GIMP test was a little low as I pointed out if you recall, so it might not be an outlier after all. You did do multiple runs to confirm, right?
    BTW: I didn't find any spelling ore grammar errors in this article, good work!
  • ballsystemlord - Thursday, July 11, 2019 - link

    Yikes, I thought higher was better. GIMP is now worse.
  • ballsystemlord - Thursday, July 11, 2019 - link

    That's confusing in this article, some benchmarks are higher is better, others higher is worse and you have to read the original article to understand which is which.
  • Ryan Smith - Thursday, July 11, 2019 - link

    Higher is better. The 3900X did 17% better with the new BIOS. (The test is a timed test, but of course we inverted this so that it could be used in those charts)
  • Andrei Frumusanu - Thursday, July 11, 2019 - link

    The charts here in this piece are showing % improvement, I took into account each test having higher/lower numbers as being the improvement.
  • ballsystemlord - Thursday, July 11, 2019 - link

    According to your picture: https://images.anandtech.com/graphs/graph14605/111...
    The new BIOS does worse on the GIMP test. The ** is marking the old result, right?
  • ballsystemlord - Thursday, July 11, 2019 - link

    3700X: 3.38 old. 3.50 new. Lower is better.
  • ballsystemlord - Friday, July 12, 2019 - link

    No, wait, your talking the 3900X only. Ok. My bad. I understand now.
  • Arnulf - Friday, July 12, 2019 - link

    "Frimware" x2.
  • serjrps - Thursday, July 11, 2019 - link

    Slightly off topic, but will you be reviewing the rest of the Zen 2 lineup? Would be quite interesting to have the full picture here in Anandtech.
  • Ryan Smith - Thursday, July 11, 2019 - link

    With the caveat that the lineup is so large that we may not review quite literally all of it, the plan is to get some more CPUs in sooner than later.
  • RSAUser - Wednesday, July 17, 2019 - link

    Can you just pick the most popular?
    Normal line was the 2600, 2700X, then lower end the 2400G (have one, bit biased) and if even more time the 2200G.
    New line it seems the most popular are: 3600 (I have one, so little biased) and then far behind the 3700X.
    Of the APU: 3400G.
  • GreenReaper - Thursday, July 11, 2019 - link

    Thanks for your update. But I hope you will also be testing the impact of MDS/Zombieload, as here?
    https://www.phoronix.com/scan.php?page=article&...

    Windows *did* patch for these, in mid-May, but this required updated microcode for the CPUs not containing it already (probably including 9th-gen review samples, as they tend to be some of the first). BIOS patches for the majority of the motherboards you were using were available prior to review, in May/June, and thus reflects the current performance of these CPUs; but the versions you used for Intel boards was from 2018.

    For your convenience, the Fatial1ty Z370 Professional Gaming i7 (Intel 9th- and 8th-gen):
    https://www.asrock.com/MB/Intel/Fatal1ty%20Z370%20...

    The X299 OC Formula (HEDT X-series):
    https://www.asrock.com/MB/Intel/X299%20OC%20Formul...

    Sadly the GA-X170-EXTREME ECC (7th- and 6th-gen) does not have a public update:
    https://www.gigabyte.com/uk/Motherboard/GA-X170-EX...
    but maybe Gigabyte could be hassled for one.
  • Ryan Smith - Thursday, July 11, 2019 - link

    "Thanks for your update. But I hope you will also be testing the impact of MDS/Zombieload"

    Yes. That's on the big list of things to do for once all the necessary software/firmware drops into place.
  • GreenReaper - Thursday, July 11, 2019 - link

    Thank you. I do appreciate that there was a lot going on these last couple of months.

    For what it's worth, microcode can also be updated without BIOS support, using a driver like this:
    https://labs.vmware.com/flings/vmware-cpu-microcod...

    I'm unsure whether this will trigger OS mitigation in this particular case, even if installed and activated on boot, but it could presumably be checked via Get-SpeculationControlSettings:
    https://support.microsoft.com/en-gb/help/4074629/u...
  • Meteor2 - Monday, July 15, 2019 - link

    Without wanting to sound a massive Debbie Downer, I'm going to repost a comment I just made on the original review article.

    It's hard to get one's head around this, but basically: *all* the Intel benchmarks *do not* include the security patches for the MDS-class flaws. The 9000 and 8000 series tests do include the OS-side Spectre fixes, but that's it. No OS-fixes for other CPUs, and no motherboard firmware fixes for any Intel CPUs

    At the very least, all the Intel CPUs should be retested on Windows 10 1903 which has the OS-side MDS fixes.

    Most if not all the motherboards used for the Intel reviews can also have their firmware upgraded to fix Spectre and most times MDS flaws. Do it.

    This is sensible and reasonable to do: no sensible and reasonable user would leave their OS vulnerable. Maybe the motherboard, because it's a bit scary to do, but as that can be patched, it should be by reviewers.

    This would result in all the Intel scores being lower. We don't know by how much without this process actually being done. But until it is, the Intel results are invalid.
  • RSAUser - Wednesday, July 17, 2019 - link

    Most users wouldn't even have a choice, Windows would usually push chip-set updates via update center.
  • nismotigerwvu - Thursday, July 11, 2019 - link

    Just saw a minor typo,

    "We’ve since been able to re-test both the Ryzen 9 3900X as well as the Ryzen 7 3900X"

    That should read "...the Ryzen 7 3700X" right?
  • Ryan Smith - Thursday, July 11, 2019 - link

    Ah, yes. That's one more 3900X than we tested!
  • JoeyJoJo123 - Thursday, July 11, 2019 - link

    "REEEEEEE! BUT AMD IS LYING ABOUT SINGLE CORE TURBO SPEEDS! THIS DOESN'T FIX ANYTHING! IF INTEL DID THIS, IT'D BE A BIG DEAL, THANKS FOR BEING AN AMD SHILL, $N$NDTECH!"
    -- sincerely, an Intel shill internet warrior who does it FOR FREEEEEEE
  • Xyler94 - Friday, July 12, 2019 - link

    I think my brain exploded from this, x3
  • guachi - Thursday, July 11, 2019 - link

    I appreciate the work in bringing us the update. There are lots of new Ryzen 7 CPUs but a complete chart (eventually) with every processor would be awesome! Lots of work but a great way to get clicks as you add more processors and the first site to have a complete set will surely net a lot of attention.
  • guachi - Thursday, July 11, 2019 - link

    Sigh... change "Ryzen 7" to "Zen 2" as I mean the entire new lineup of Zen 2 processors.
  • binkleym - Thursday, July 11, 2019 - link

    Given that several motherboards are having to remove features (RAID, graphical BIOS, user profiles) from the BIOS in order to fit the AGESA for Zen 2, it would be nice if motherboard reviews would start mentioning the size of the BIOS, so we can easily discern which motherboards are designed to be future-proof, and which ones are nickel'd and dime'd into early obsolescence.
  • Irata - Friday, July 12, 2019 - link

    AFAIK, this has been mitigated for X570 boards but I definitely agree with you that this would be an important point and good to know.
  • Solidstate89 - Friday, July 12, 2019 - link

    None of the X570 motherboards suffer from this. Only the last generation chipsets/UEFI.
  • Arbie - Thursday, July 11, 2019 - link

    The percent change figures are interesting, but more directly useful and more appropriate would be to re-publish corrected charts from the first article!

    And / or to correct the charts therein - as those will be what people find when they search on Ryzen 3000 reviews. These differences are not insignificant.
  • Arbie - Thursday, July 11, 2019 - link

    Sorry - missed the sentence that you had corrected the original article. Thanks.
  • boozed - Thursday, July 11, 2019 - link

    "Improvement difference"

    Either one of those words would have worked just fine on its own.
  • Assimilator87 - Thursday, July 11, 2019 - link

    Should have been "AMD-review-athlon" =P
  • azrael- - Friday, July 12, 2019 - link

    Does any of this impact Ryzen 3000 on an X470 board? Or perhaps more to the point: Is this something that would benefit X470 boards as well, given a new BIOS update?

    It's obvious that AMD and vendors would like to push the new X570 boards, so perhaps hoping for an update with these fixes/fine-tunings is moot.
  • Qasar - Friday, July 12, 2019 - link

    i think for the ryzen 3000 series to work " properly" x470 boards would need a bios update.. by " properly " i mean, recognized and to have the features of the new chip, to work.. but it should work with out one.. just maybe not very well...
  • Death666Angel - Friday, July 12, 2019 - link

    The "fix" was to a broken system. So if the 4xx and 3xx chipsets get the appropriate behavior BIOS implementation, then it should function as it does here. But that is up to the individual motherboard vendors. Most upper end motherboards should get decent to good support (4xx starting at $100). 3xx might be more limited to the higher end SKUs.
  • SydneyBlue120d - Friday, July 12, 2019 - link

    Do You plan to review the new APU too?
  • BPB - Friday, July 12, 2019 - link

    I think there are a lot of folks waiting to see how the new APU does. I believe that comes out later this quarter, doesn't it?
  • GreenReaper - Friday, July 12, 2019 - link

    They're doing a lot to try to make it appealing. It's still Zen+ though - and Vega, not Navi. I'll be more interested once they shrink Zen2 down and maybe slot AV1 support in. Without that it seems unlikely to deliver competitive performance to power usage. If nothing else, they might be able to toss more cache in - or make it smaller, and hence potentially even cheaper.
  • RSAUser - Wednesday, July 17, 2019 - link

    And it's also still on the older 14nm node, so not even close to as power efficient. There will probably be quite large APU jumps in the 4000G series.
  • urbanman2004 - Friday, July 12, 2019 - link

    3700X FTW
  • AGS3 - Friday, July 12, 2019 - link

    Ryzen 3000: Asus opens up PCIe 4.0 support for selected X470 and B450 boards
    https://www.guru3d.com/news-story/ryzen-3000-asus-...
  • Qasar - Friday, July 12, 2019 - link

    not according to amd ( again ) : https://www.anandtech.com/show/14639/no-amd-still-...
  • peevee - Friday, July 12, 2019 - link

    "We’ve since been able to re-test both the Ryzen 9 3900X as well as the Ryzen 7 3900X"

    Really?
  • znd125 - Friday, July 12, 2019 - link

    Curious what were the other 3 "launch boards"?
  • Maxiking - Saturday, July 13, 2019 - link

    So fraund confirmed, no 4.6ghz boost on a single core as advertised.
  • Korguz - Saturday, July 13, 2019 - link

    you STILL on this bs ??? come on.. its called teething issues on a new platform, with a new cpu.. it happens.. even you mighty intel has had issues with its new platformd over the years.. drop this crap already...
  • Qasar - Saturday, July 13, 2019 - link

    yea.. he should drop this... geeze
  • Maxiking - Sunday, July 14, 2019 - link

    No, it is called false advertising. As I said, a large number of those cpus can not reach 4.6ghz on a single core because the silicone quality is poor.

    They specifically chose to use such boosts /12core, 16 core - 4.6 and 4.7 ghz/ to confuse people, so they could assume the cpus would be able to reach the frequency on all cores after overclocking. The fact is the node is so poor, it barely reaches 4.3 ghz on a few chips and often with voltage over the limit.

    Recently AMD claimed that had jebaited Nvidia. It is ridiculous when you think about it. They use a 7nm node yet their architecture is so terrible, they can't beat a gpu from 2016, 1080ti and and competing with mid range Nvidia cards.

    So yeah, this is the one and only successful jebait from AMD, the false boosts.
  • Korguz - Sunday, July 14, 2019 - link

    then WHERE is the proof ?? and i dont mean forum posts by random people. i mean actual articles on reputable sites that are showing this. which so far, you have shown NONE, there fore, YOU are the FRAUD, and are just trying to spread BS and are tolling. either post links to REPUTABLE sites, or shut up and drop this.
  • Maxiking - Sunday, July 14, 2019 - link

    So this site isn't reputable?

    I already mentioned plenty of sites which weren't able to fully utilize the boost freq. Yet I am always being told by the fanboys that bios patches will change that. So far, a second test on this site and proving my words, again no boost to 4.6ghz on a single core. There is a graph. Look at it.

    You know, being a fanboy doesn't make things better, companies don't care about. You sure know that, right.
  • Korguz - Sunday, July 14, 2019 - link

    this site is reputable, and as i said before, teething issues with a new platform, and with an updated bios, even your mighty intel has has the same types of issues over the years and issues look to be fixed, and i bet, after a few more updates, things will get better, hardly any reason to accuse a company for fraud or miss advertising, but by all means, file a law suit against amd for this, lets see how far you get.

    the only thing you mentioned.. were FORUMS, not articles posted by the sites operators, BIG difference, a forum, is random people, who are usually not part of the sites staff, vs an article published on the site, written by someone who runs the site. Maxiking, YOU seem to be the ONLY only one complaining on this site about this. you want to call me a fanboy, go ahead, but i hope YOU know, that you also come across to me, as some one who wants to make a big deal out of something, where there is NO REASON TO. if amd did not do anything to fix this supposed issue in a timely matter, then maybe there is a reason to cry fraud and miss advertising, but it looks like they are taking steps to fix this as soon as they are able to find and implement a fix
  • Maxiking - Monday, July 15, 2019 - link

    You are a typical fanboy, I didn't mention Intel once unless you tried it to force into the discussion.

    So let me tell you it again, Intel doesn't have such issues and fraud marketing. If they mention turbo and its frequency, a single core under a single threaded task ALWAYS RUNS at such frequency.

    It is different with AMD, because the node they use is terrible and in order to gain the boost clocks they have to push voltage over the safety limits 1.4 - 1.5 , that's why they boost in the short spikes I already described.

    I mentioned forums and sites like techspot, gamer nexus and so on.

    I am not gonna sue AMD, I don't buy their products, be it GPUs or CPUs.

    You are a fanboy, you acknowledged the issue yet you try to soften it and blame the bios. The opposite is true, 4.6 GHz boost is a paper dragon and a big number of CPUs can't reach it because of the poor node.

    This is a second test, it confirmed my words, nothing changed, period. Like always with AMD product, there always updates, be it Windows updates, scheduler updates, bios updates which are supposed to fix issues. And as always, those updates have fixed and will fix NOTHING. I wonder how dumb you have to be to believe this story again and again. Anyway, Amd confirmed my words on twitter in PR language and admitted to the fraud by saying that boost clock are not guaranteed even on a single core. So yeah, fraud. Bye
  • Retrofire - Monday, July 15, 2019 - link

    I'm pretty mercenary on my hardware purchases, but I did violate my own rule and early-adopt a 3900X.

    I can say that my experiences with this AMD part align with your narrative. Of my 12 cores, six max out at 4.4ghz, four max at 4.3ghz, and two can only sustain 4.2ghz. Using PBO I have *never* see a boost to 4.6ghz, and *may* have hit 4.5ghz for a split second once.

    This isn't due to running a stock cooler, or gimped BIOS, or lagging chipset drivers. There is no way any of my cores can sustain even 4.5ghz at safe voltage, much less 4.6ghz and you'll be able to knock me over with a feather if my intuition is proven wrong here and some magic bullet AGESA rev drops that changes this.

    That said, I'm still very impressed with the chip overall. No, it doesn't overclock. No it doesn't quite hit its advertised specs, which will be a problem for AMD if they cannot engineer around the silicon and there's a class action.

    At max clocks on all cores, I can hit a 7500 Cinebench R20 consistently at a safe voltage, and for multi-core that's nothing short of stunning out of a $500 part. I bench about 6% below a stock 9900K single-core, which doesn't bother me as I never game at low-resolution, and I'm GPU-bound 99% of the time anyway.

    While the name-calling ultimately weakens his argument, Maxiking isn't wrong: It seems clear to me that AMD misrepresented the specifications around their new chip. I'll gladly take the performance gain and eat my words if they're able to fix this with microcode and drivers, because after four days of non-stop tweaking, there just doesn't appear to be any more headroom left on these parts to eke performance out of.

    I would say this with much more chagrin if the 3900X wasn't performing as well as it does at the clocks it *can* hit.

    Overall, this feels (initially) like an unforced error from AMD. They could have had all of the glory and much less of the negativity if they had lowballed their clock specs and provided room for overclocking.
  • Korguz - Tuesday, July 16, 2019 - link

    um, intel doesnt have issues with fraud ??? um hello ? have you NOT sees the reports of intels use of TDP ? how they " claim " some their cpus use 95 watts, but in reality use up to 200 watts ? BTW, how is that intel cpu that you use ( as you state you dont use amd's products, so you must use intel ) i bet its probably not uses the watts intel " claims " i wonder how dumb YOU have to be to keep believing the lies and BS intel feeds you. so yea, NOT using the power your products claim, but using A LOT more, fraud also. good riddance
  • Maxiking - Monday, July 22, 2019 - link

    You are uneducated, TDP doesn't mean power consumption but the amount of heat dissipated.
  • Qasar - Tuesday, July 16, 2019 - link

    um maxiking, why is it others have questioned you on this as well, but you ONLY seem to responding to Korguz ??? also WHY does it also seem like you don't acknowledge the negative aspects of intel ? you ONLY seem to be going after amd for faud, but you don't do the same for intel ? alufan pretty much said the faud that intel has done over the years, but you dont seem to see that ? the biggest one, seems to the the claimed watts their cpu's use, so i hope you have a cooler on your intel cpu, that can handle 30-50 more watts then the chip is rated for, you'll need it. seems to me.. you are NO better then he is, that sounds like and intel fan boy to me.
  • Qasar - Sunday, July 21, 2019 - link

    hmm no reply from maxiking, thought he wouldnt reply to that question. hes fine with accusing amd of fraud, but to accuse intel of the same, he wont do. figures
  • Maxiking - Monday, July 22, 2019 - link

    So there is my comment. You are uneducated, TDP doesn't mean power consumption but the amount of heat dissipated, it informs you how much of heat the cooler must be able to dissipate in order to keep the cpu cool enough to run.

    Get it? 1700x TDP was 95W yet there were tasks it managed to consume 120 or even 140w on stock settings. Like do you even watch reviews? It was the same with 2700w.

    but mimimimimimi AMD good mimimimimi Intel bad
  • Maxiking - Monday, July 22, 2019 - link

    You are uneducated, TDP doesn't mean power consumption but the amount of heat dissipated.
  • Xyler94 - Friday, July 26, 2019 - link

    TDP stands for Thermal Design Power. Intel misleads consumers, in that they use Base Clock to measure TDP. Meaning, if your cooler can only dissipate 95W, good luck going anywhere above the base clock. If you happen to have a Z370 board that goes over the limit, the chip can output as much as 180 to 220W of heat, for a chip that is rated at 95W. And if that isn't cause for fraud, then I don't know what is.

    But you know why it isn't fraud? Neither Intel or AMD guarantee boost clocks. Intel lists Single Core boosts for their CPUs (except one, 9900KS, but that's a weird case). AMD does the same, single core boost. But Single Core Boost means a single core is being used ONLY. If the OS wants to use a core while the single core is being used, then it'll drop speeds.

    Unlike AMD's Zen2, Intel's cores can sometimes run 2 cores are the rated max speed, but anything above that, it falls short. Zen2 is pushed so hard, that to get the rated single core boost, you need to get 100% single core utilization, all other cores need to be at 0%.

    Honestly, you keep calling others uneducated, but haven't shown how educated you are in the subject.
  • RSAUser - Wednesday, July 17, 2019 - link

    I'm running a 3600 as of today.
    I'm getting boosts of 4.6 on a single core in cinebench after I upoped the power limit a little, so seems to be running fine. All core for sustained (was encoding some video) I was getting around 4.1GHz at about 75W, 1.3V.
  • 29a - Monday, July 15, 2019 - link

    I'm pretty sure that the base frequency is the only guaranteed frequency.
  • alufan - Monday, July 15, 2019 - link

    Maxking, you want fraud ask intel where the 10nm they promised not only the public but their own shareholders you want fraud ask intel about the TDP value on its chips and fwiw I have a 3900 that does hit its 4.6 on a couple of cores in cycles yes one does it consistantly others do it more in sequence, my GPU is a 2080ti so I game at high res so frankly any intel chip even close to the performance of my AMD will be a hell of a lot more expensive and thats not without factoring in the huge extra amount of power it takes to run, because of the TDP, the AMD also gets a cooler included which actually works to cool the chip intel?
    It matters not what we as individuals think but the truth is intel has held us all ransom for many years since the core 2 came out at low core counts and old fashioned and outdated insecure products, we should all be grateful that AMD has come straight back in within 2 years and kicked butt otherwise you would still be handing over £600 for a 4-6 core CPU with security issues
  • gamoniac - Saturday, July 13, 2019 - link

    Andrei and team,
    Thanks for taking the time to write a separate summary article other than updating the original article. Since I spent so much time reading the original review, this summary is much appreciated.
  • rower30 - Tuesday, July 16, 2019 - link

    My AMD Ryzen 5 2600X clocks to over 5 GHz average (5030, 5030, 5030, 5030, 4999, 5062 MHz) on all six cores with an ASRock Taichi X470 using DDR4 G.SKILL XMP-3200 14 14 14 34 48 RAM and the latest 3.43 BIOS using CPUID HWMonitor Version 1.40.0. BIOS is set to "auto" overclocking. Max core is shown as 1.463 Volts with a CPU temp of 69C using a Noctua NH-D15S cooler. Room is 82F. This playing QUAKE 2XP most recent graphical mod edition.

    I can loop Heaven Benchmark 4.0 and the boost numbers are the same. The 2600X seems to be what it is supposed to be, and runs better than any PC I've ever built. I use an ADATA NVme SX8200NP 1TB drive as a boot and applications drive, but this won't impact CPU boost properties.

    If my 2600X can do this, I have a hunch that the 7nm replacement can do it with the right BIOS. I really, really don't think that the die shrink LOST boost frequency. Maybe not gained but lost? I see a just over 4.5 GHz steady boost in the graph, so if there is a tardy 100 MHz, it is few in the early stages of this design.

    I understand the the point to the argument of boost clock being shy of spec but the CPU's IPC throughput is vastly better so the CPU is far better. It uses less power by far, runs faster EVERYWHERE and is expensive but expensive like it SHOULD be. I feel I get what I paid for, but was not ripped off like, um, that other CPU. If I built today, I would get a 3700X without a thought to Intel.

    And, my AMD X470 will go to a 3900X as it's last stop before the next generation will use PCIe 4.0 right (cooler running) and even have proper NVMe drives. The next CPU after the 3000 series will likely not be useful on X470.
  • rower30 - Friday, August 2, 2019 - link

    Correction! My CPUID HW 1.40.0 monitor program boost frequency is WRONG! I downloaded AMD Ryzen Master and also HWinfo64 and BOTH of these are spot-on to each other with boost frequency, my 2600X boosts to 4.2 GHz on all cores. Why the CPUID HW program consistently shows so much higher is a mystery to me, but it is wrong. Watch out with this program on boost frequency. So the 2600X meets the boost on all cores, where the newer 3000 series design meets it on a strong single thread core. My take is this, the 3600X is a far better CPU than my 2600X. Do I care that it does not boost simultaneously or as high on as many cores? No, I care about how it actually runs a program, and it is ALWAYS better than my 2600X. Would you buy a 2600X verses a 3600X building a new PC? No, you would not. They are priced the same MSRP at their introduction and the 3600X is a big step up for what you get over when I purchased the 2600X a year ago. The 2600X is a good deal a hundred bucks less than the 3600X today, though.

    I actually turn off SMT playing older games like QUAKE and QUAKE 2 as old games don't seem to like virtual cores and freeze a half second or so randomly, and stutter if you clip near walls. Using only logical cores fixes all that. Six real cores is plenty. I do nothing on the desktop that needs 12 cores.

    Sorry about posting wrong info everyone.
  • PProchnow - Saturday, July 20, 2019 - link

    I saved my GeekBenchs and here is the best, Hopin "A" allows HTML links.

    https://browser.geekbench.com/v4/cpu/13906968

    I feeel it is something like the biggest processing deal available, maybe of all times all things factored in. A mere upgrade for me, not a total rig. My Ryzen 7 2700 now seems nearly shamefully weak althought it replaces the 2200G in the HTPC
  • dad_at - Sunday, July 21, 2019 - link

    Thanks for the article and this update. I noticed there is no "chromium compile" test with the new ryzens? Why? I assume 3900X should perform pretty strong in the code compilation tasks compared to intel CPU...
  • Tech-fan - Sunday, July 21, 2019 - link

    I find it desterbing that reviewers get all the best boards to test with the new CPU's. The majority of gamers have not that much cash for a MoBo, so why not testing with low end boards of the chipset? That is what most gamers will buy for their own build.
  • Maxiking - Monday, July 22, 2019 - link

    AMD fraund getting finally the attention it deserves

    https://www.youtube.com/watch?v=x03FyPQ3a3E

    check at 05m25s
  • markoyelo - Wednesday, July 24, 2019 - link

    just found a detailed review on AMD Ryzen Gen 3 processor https://telegraphstar.com/amd-ryzen-gen-3-processo...
  • mblataric - Wednesday, August 7, 2019 - link

    It would interesting to see how PB2 on X570 compares to PB1 on X470 boards.
    I am strongly considering X470 since I do not need PCIE4 and idea of having 6000rpm fan on south bridge brings painful, head-aching, ear-ringing memories :-)

Log in

Don't have an account? Sign up now