GP discusses the latest CPUs & GPUs

Thread Rating: 5 votes, 5.00 average.
(5 votes)
Results 15,376 to 15,400 of 15535

  1. Post


    not sure why it would work with Ryzen? , although its CL 16, I think its fine

    e: oh lmao listed for Intel, thats BS tho

  2. Post
    Are you applying a discount or something? Showing up quite a bit higher in price for me.

  3. Post
    Good deal
    DW wrote:
    Are you applying a discount or something? Showing up quite a bit higher in price for me.
    Is yours also MicroCart as the seller?

  4. Post
    Oh. That isn't shipped to NZ (MicroCart don't ship to NZ), would need to go via YouShop or similar.

  5. Post
    I bought it...felt like buying something, this will also get me started on my 3600 build, have to finally say goodbye to my 2500k

    https://www.amazon.com/gp/offer-list...&condition=new

    we'll see how it goes, only comes in August

  6. Post
    SirGrim wrote:
    Just watched the latest video

    Will that be the last on YouTube?

    He puts out really good analysis. I found his theory quite interesting as well - he seems to believe that ryzen 3000 is low quality parts that AMD is overpricing to fleece customers on the back of having no competition and AMD is saving the good 7nm chips to sell later on.

    But then he goes on to say he will no longer cover AMD products because they lied on a couple slides about Zen 2, which seems like an over-reaction. Then he says he has other reasons to find out what those are, you can subscribe to his patreon.. lol
    OMG, he is such a drama queen.

    He kind of has to say they are selling low quality stock just to back up his leaks, and of course no company is going to comment on leaks.

    And you price your product based on the market. If AMD could sell them for more they would, if Intel were to come out with a new cpu tomorrow that is much better than AMD, AMD would have to adjust their prices.

  7. Post
    You don't think desktop gets the junk silicon? Certainly makes sense to me. You certainly won't find the junk stuff selling for $000's as a server chip. Makes sense that the more you pay, the better quality product you get.

    We are just lucky that even the worse silicon they can sell, performs really well compared to previous chips. 4000 series will just be a refresh with a more mature silicon. If all silicon was the same, the lower cored chips would clock to the exact same frequency as the higher cored ones. Four core and six cores are more than likely chips that had defects in cores. That's how it crumbles.

    Selling the lesser performing stuff keeps overall pricing down too, which is good for us. What we do need now is Intel to create some competition to put AMD under pressure to deliver closer to the best they can. For us bottom feeding desktop users anyway.

    Will be interesting to see how Threadripper performs on Zen 2. Especially clock wise.

  8. Post
    Junks also a relative term. Junk compared to higher binned Zen 2 chiplets sure but it's still better price to performance than other desktop CPUs you can buy today.

  9. Post
    Relative for sure. How often do you see 9900k chips that don't hit the mark though? Intel are evil, but they do seem to keep the chips they sell performing at least where they claim they will.

    Meanwhile AMD's chips aren't hitting advertised boost levels, except in rare situations for minimal amounts of time. Either telling porkies or sell off chips that don't cut the mustard. I think they claimed 3.5or6ghz chips will boost to 3.7, 3.75ghz under the right conditions. I've yet to see anyone claim to have got that much of a boost.

    I think Ryzen 3000 is great, but that has to leave a sour taste when the figures they supply don't match the real world results.

    Just my2c anyway, as a fan of AMD products.

  10. Post
    Fragluton wrote:
    You don't think desktop gets the junk silicon? Certainly makes sense to me. You certainly won't find the junk stuff selling for $000's as a server chip. Makes sense that the more you pay, the better quality product you get.
    Yes they do but not at the difference he is talking about - more than 10% difference in clock speeds. Have a look at EPYC vs RYZEN/Threadripper first gen clock speeds and TDP, there is not that big a difference.

  11. Post
    Kog wrote:
    I bought it...felt like buying something, this will also get me started on my 3600 build, have to finally say goodbye to my 2500k

    https://www.amazon.com/gp/offer-list...&condition=new

    we'll see how it goes, only comes in August
    if it arrives... my 32gb 3200mhz kit never did

  12. Post
    Fragluton wrote:
    Relative for sure. How often do you see 9900k chips that don't hit the mark though? Intel are evil, but they do seem to keep the chips they sell performing at least where they claim they will.

    Meanwhile AMD's chips aren't hitting advertised boost levels, except in rare situations for minimal amounts of time. Either telling porkies or sell off chips that don't cut the mustard.

    I think Ryzen 3000 is great, but that has to leave a sour taste when the figures they supply don't match the real world results.

    Just my2c anyway, as a fan of AMD products.
    Yeah, the boost clocks advertised seem best case scenario - similar how for Navi they advertised boost clocks and game clocks where the game clocks are a more realistic target.

  13. Post
    Pxndx wrote:
    if it arrives... my 32gb 3200mhz kit never did
    i think ure being a bit dramatic. Its still on the way

  14. Post
    Intel is sending out 10 core Comet Lake (14nm+++) and other Comet Lake CPU's to motherboard vendors for qualification and testing

    https://www.pcgamesn.com/intel/comet...-qualification

    Expected launch is October

    What the hell happened to the 9900KS, MIA lol?

    Bullion wrote:
    Yeah, the boost clocks advertised seem best case scenario - similar how for Navi they advertised boost clocks and game clocks where the game clocks are a more realistic target.
    But with Navi you can actually hit the boost clock 24/7 by overclocking and better cooling. With Ryzen 3000, well yeah not really

  15. Post
    Bullion wrote:
    Yeah, the boost clocks advertised seem best case scenario - similar how for Navi they advertised boost clocks and game clocks where the game clocks are a more realistic target.
    Watch this and tell me if you think it's accurate. Reviewers are using top spec boards, but I don't think anyone has got near a boost like it. Certainly not lifting the whole curve up for the whole time while gaming, which is implied. Netting a smaller boost, on a single core, for half a second, once in a whole test is a bit of a laugh. Which is what GN seemed to find. The actual highest boost was for a split second and not seen again. I just think it's misleading, and heading the way of Intel / Nvidia in trustworthy stakes. They had no reason to lie about anything, the chips are great. Yet they still did.


  16. Post
    GN covers it, if you get bored lol tl;dw - PBO doesn't do bugger all

  17. Post
    Fragluton wrote:
    Watch this and tell me if you think it's accurate. Reviewers are using top spec boards, but I don't think anyone has got near a boost like it. Certainly not lifting the whole curve up for the whole time while gaming, which is implied. Netting a smaller boost, on a single core, for half a second, once in a whole test is a bit of a laugh. Which is what GN seemed to find. The actual highest boost was for a split second and not seen again. I just think it's misleading, and heading the way of Intel / Nvidia in trustworthy stakes. They had no reason to lie about anything, the chips are great. Yet they still did.

    One thing is there no mention of all core boosts and single core boosts. Intel stock advertised boosts are 1 (or 2) cores, though many able to overclock all cores to around the boost frequency. AMD has never really been like that, 2700x boost was 4.35 (if conditions were right) and many reviews said don't overclock because you can't reach that on all cores and you can get worse performance in games because for many games frequency > 8 cores - hence some people started trying BCLK overclocking.

    This is what AMD supplied to the reviewers:
    Name:  D-6g4r8U8AACahr.jpg
Views: 124
Size:  48.4 KB

    From Hardware Unboxed
    https://twitter.com/HardwareUnboxed/...31949723729920
    Didn't run into this problem. Pretty sure our 3900X hit 4.6 GHz single thread as expected. It doesn't clock to 4.6 GHz all-core or even with two threads. Here's AMD's clock speed chart from the review guide:

  18. Post
    All said and done I don't regret getting the 3900x. All the cores sit at 4.3 all day and nothing seems to slow it down. Would be nice for my ocd if I ever saw at least 1 core hit 4.6 but I don't think I'd ever actually notice the difference. It could also be lacking due to it being brand new architecture and bios and microcode being immature.

  19. Post
    I feel like they should have just released without PBO then. Since it doesn't actually work, or offer 200mhz in any situation on any current chips.

    Yeah like I say, the chips are great. I just feel like they tried to add too many acronyms on to the end to sound cool. But they just don't actually do anything. The whole PBO video should just not exist. No one with a single core boost near the 4.55ghz mark mentioned, has ever seen 4.75ghz.

    Either the chips aren't of high quality enough to see the PBO gains mentioned
    PBO is broken
    PBO doesn't actually exist

    lol

  20. Post
    pyronical wrote:
    So generally the consensus is that with PBO on, you should get to 4.2-4.3Ghz on a 3700x.

    I am getting 4.4Ghz pretty consistently - did I win the lottery?

    I've done this again, played about 30mins of Civ6 and I got all 8 cores to show 4399. During stress, they were actually sitting at 4349 or 4399 for periods of time after all. Did I just get lucky or is it only the 3900X with this issue?

  21. Post
    People rush to defend AMD because they are seen as the good guys. The plucky underdogs out to put up a good fight against the Intel Evil Empire. The reality is AMD were very disingenuous about their boost clocks. I have a 3900x. It literally says on the box boosting to 4.6. Now, no one expected all 12 cores to boost to this speed but I have yet to see a single core on my cpu boost to that speed. 4.496 is the max I have seen. Don't get me wrong, I am happy with the CPU and it's a beast and deserving of the praise it gets, but call a spade a spade. AMD were sneaky.

  22. Post
    pyronical wrote:
    I've done this again, played about 30mins of Civ6 and I got all 8 cores to show 4399. During stress, they were actually sitting at 4349 or 4399 for periods of time after all. Did I just get lucky or is it only the 3900X with this issue?
    How often do you see the 4.6Ghz boost? I imagine you do have a good chip, though I haven't studied what other chips all core boosts are. Would expect to see some max boosts at the 4.6ghz mark if it was performing as AMD say it should. Mix some single thread workloads in and see if any cores change their max frequency to something higher.

    What frequency are the cores at when all being utilised?

  23. Post
    Fragluton wrote:
    How often do you see the 4.6Ghz boost? I imagine you do have a good chip, though I haven't studied what other chips all core boosts are. Would expect to see some max boosts at the 4.6ghz mark if it was performing as AMD say it should. Mix some single thread workloads in and see if any cores change their max frequency to something higher.
    4.4 you mean (3700x), pretty often, not constant like you'd see on an intel chip. But it gets there when there's something to compute, so I dunno - I'm happy.

  24. Post
    He has a 3700x so it won't boost above 4.4. I am not surprised at 4.4, my 3900x gets those speeds on the cores from time to time. 4.6 is another matter.

  25. Post
    pyronical wrote:
    I've done this again, played about 30mins of Civ6 and I got all 8 cores to show 4399. During stress, they were actually sitting at 4349 or 4399 for periods of time after all. Did I just get lucky or is it only the 3900X with this issue?
    Is the 3700x one chiplet or two? I'm sure there was a video floating about where it was claimed the 3900x has one high binned chiplet and lower binned chiplet, might well explain why 3900x can't go any faster than 3700x even though the box says differently. If they fix whatever needs to be fixed to force the single threaded apps to the better chiplet we might actually see our 4.6 boost.