Page 14 of 14 FirstFirst ...
4
12
13
14
  1. #261
    Quote Originally Posted by mrgreenthump View Post
    That said, Radeon VII does have it's place and is undercutting Nvidia by a lot in that market. Just gaming isn't it and they should've never marketed this as a gaming card.
    ... it performs the same as a 2080 and costs the same as a 2080. It isn’t undercutting anything.

    Also on CPU side, what are you expecting? Ryzen 1 undercut Intel by 50%. Ryzen 2 undercut their own line by 35%. Be as mad as you want towards Intel, but as far as CPU pricing goes.. AMD has been doing a quite decent job keeping them low.
    AMD is accepting nearly no profit margin to gain market share, which is why they are so cheap. Not that that is a bad strategy or anything.

    - - - Updated - - -

    Quote Originally Posted by mmoc8f3638b110 View Post
    To assume for whatever reason that these companies act honestly and adhere to whatever market rules that have been invented out of thin air is ludicrous. Profits from these corps and especially the tech industry come from two main sources, cheap labour and huge government subsidies.
    Neither Intel or AMD gets any major subsidies from the US government. Intel makes almost all of its CPUs here in the US, so it isn’t “cheap labor”. Not that there is a lot of labor involved. The scale being worked with, people aren’t involved except as monitors.

    Things might well be as you say in regards to the cost of building a high end card,
    No, what i said was that ONLY the high end cards have OK margins - and only for nVidia. Like.. the 1080Ti and the Titan Xp, and the 2080Ti and maybe the 2080 have OK margins (above 10%). Lower end cards have even LESS margin. They make money purely on volume.

    but that's not my point - prices are high at the top end to drive up prices for the mainstream tier cards.
    Except they dont. The cost of the card is almost entirely determined by the cost of the parts and labor to assemble it. You can price the parts yourself, exept for the actual GPU die itself (but we have an idea of what those cost, as if YOU were to order something similar from Global Foundries, youd know what it was going to cost). A GTX 1060, for instance, had about a 7% margin for nVidia (the founders edition). AIB partners make even less (as they have to buy the GPUs from nVidia, instead of getting them at cost from GF or TSMC) - i remember it was EVGA that was telling Paul and Kyle (BitWit) that their average margin is sub-5%.

    If anything it shows more so how these corps collude with each other to maintain profits.
    How are they colluding together? Theyre selling them barely above what it costs them to make them.

    I mean it just so happens that Nvidia charges 699 MSRP - higher than something that performs practically the same last gen
    Uhh.. wrong? The 2080 performs as good as the 1080Ti or slightly better... and is the exact same price. Both are 699MSRP.

    - and it just so happens that AMD can sell their cut down silicon at a small profit margin at exactly the same price as their competitors third tier flagship card.
    Second-tier. The Titan is not part of the consumer product stack. And the main reason they cant sell that card for cheaper is the HBM2. If they had gone with GDDR6 they could most likely have sold that card 200$ cheaper.

    My point remains, prices creep up for the mainstream consumer for mid-tier products e.g. the 2060s and whatever iteration AMD comes out with and Halo products that get released, such as the Titans and Radeon VIIs ensure that cards lower down the product stack remain high.
    You dont have a relevant point. The 2060, btw, is 30$ CHEAPER than the previous generation product in that place in the stack. (It is NOT the replacement for the mid-range 1060, as with the removal of Titan, everything has shifted UP on product category - it is the replacement for the upper-mid-range 1070). The 1070, which it is actually the replacement for in the stack, was 30$ more expensive (380$ vs 350$) at launch.

    Anyone with any memory of GPU and CPU pricing can see that prices are creeping up for the mainstream.
    Only if they aren’t intelligent enough to notice that because of the removal of Titan, the previous naming scheme and how it relates to where a product sits in the stack is no longer valid.

  2. #262
    Now if only they could be better than intel chips.

  3. #263
    Quote Originally Posted by Kagthul View Post
    ... it performs the same as a 2080 and costs the same as a 2080. It isn’t undercutting anything.
    6.9 TF of FP64 for $699 vs 7.4 TF of FP64 for $3000.. Hmm I do wonder.
    Last edited by mrgreenthump; 2019-01-12 at 11:36 PM.

  4. #264
    Quote Originally Posted by Axxil View Post
    Now if only they could be better than intel chips.
    Assuming they are making any kind of progress with zen 2 at all it will not really take much. They are in spitting distance of intel's IPC, all that was really holding AMD back this go around was really poor overclocking ability.
    The whole problem with the world is that fools and fanatics are always so certain of themselves, but wiser people so full of doubts.

  5. #265
    Quote Originally Posted by mrgreenthump View Post
    6.9 TF of FP64 for $699 vs 7.4 TF of FP64 for $3000.. Hmm I do wonder.
    You know that the pro divers wont work with the Radeon VII, right?

    If you buy one thinking you're going to use it as a workstation card, you're going to be in for a nasty surprise.

  6. #266
    I fixed the gpu board stress/Flex I know...was using the outer holes...you know that warps stuff when it's tight....Pretty even now and think even better seat and cooler...it's like a 290 block anyways...universal/slotted...copper/Acrylic. There's some trash/corrosion forming in that block too....

    - - - Updated - - -

    [IMG][/IMG]


    That's much better lol
    Last edited by JRZoid; 2019-01-13 at 01:35 AM.

  7. #267
    So, couple of things... I will list them in Bullet point.

    1- AMD stated that they were using an engineering sample, which means it is not the final product. Assume that IPC and frequencies WILL go up (my guess is 5 to 12% from current). And they did this using double digit reduced power levels... On an engineering sample. I am gonna guess they used R5, but will have to wait and see.

    2- AMD stated later that Navi is set to be the gaming GPU that will come out much later this year. The idea of Radeon VII being a gaming card is probably meant to illustrate that the card is not JUST a workstation card.. but we all know what a card packing 16gb HBM2 Ram is gonna be targeted at. I think this was a fumble on AMD's part. They should have either announced Navi as the gaming project, or focused on the workstation prowess of RVII.

    3- AMD used an engineering sample and BEAT an all ready shipped flagship CPU. Intel doesn't have anything major set to release before 2020. Maybe it is 10nm, maybe it is 7nm. But they are, as of right now, behind the curve.

    4- Intel may have shot themselves in the foot by not having a response ready for R2.0 performance. Nvidia was quick to jump out and decry the RVII as terrible. Whether right or wrong, it sends a message of confidence. Intel failed to bring that.

    5- I will reserve judgement on the fate of AMD gaming GPU capability until we learn about the fate of so called project "Navi". AMD did state later that they have their eye on Ray Tracing and will have their own version. They said it was not a priority for them to have it because the eco-system has not arrived to put it to good use big picture. I agree with this. I give fuck-all about Ray Tracing, and for the most part DLSS. I want 4k, high FPS. AMD is set to deliver that.
    "When you build it, you love it!"

  8. #268
    Press F for the pack of Camels.

    Oh noes, it's still there.

  9. #269
    Quote Originally Posted by CryotriX View Post
    Press F for the pack of Camels.

    Oh noes, it's still there.
    And you can clearly see the PSU is just.... full of dust. Add that to the open case that's not made to be open, the mess of power cables when you can easily see that it at least has HOLES for cable routing... such a disaster lol

    Actually, it's probably not just dust, judging by the large amount of cigarette butts in the other picture....

    Don't smoke next your PC folks.

  10. #270
    nooo...that's the Beige color of the Noctua Fan.

    I dumped the Loon out of it and adjust the pot..sitting right at about 12.1v Perfect...

  11. #271
    Holy Priest Saphyron's Avatar
    Join Date
    Feb 2012
    Location
    Netherlight Temple
    Posts
    3,185
    Quote Originally Posted by JRZoid View Post
    lol
    OMG kill it! Kill it with fire before it lays eggs.
    My case was horrible with cable management (can't even close the side of it properly since space is so tight) (buying a new one soonish)
    And even mine looks like a Picasso painting compared to that.

    An old picture, I have replaced GPU since then.
    Last edited by Saphyron; 2019-01-13 at 03:23 AM.
    Active Wow Player Raider.IO | Semi-active D3 Player | Permanent Retired EVE Player | Inactive Wot Player | Retired Openraid Raid Leader| Inactive Overwatch Player | Semiactive HotS player | Youtube / Twitter | Steam | My Setup

  12. #272
    Quote Originally Posted by Kagthul View Post
    You know that the pro divers wont work with the Radeon VII, right?

    If you buy one thinking you're going to use it as a workstation card, you're going to be in for a nasty surprise.
    I would think that is obvious considering it ain't called Radeon Pro.

    That said, you can use pretty much any Radeon(some oldies not supported) with ROCm.

  13. #273
    Deleted
    Quote Originally Posted by Kagthul View Post
    ... it performs the same as a 2080 and costs the same as a 2080. It isn’t undercutting anything.



    AMD is accepting nearly no profit margin to gain market share, which is why they are so cheap. Not that that is a bad strategy or anything.

    - - - Updated - - -



    Neither Intel or AMD gets any major subsidies from the US government. Intel makes almost all of its CPUs here in the US, so it isn’t “cheap labor”. Not that there is a lot of labor involved. The scale being worked with, people aren’t involved except as monitors.



    No, what i said was that ONLY the high end cards have OK margins - and only for nVidia. Like.. the 1080Ti and the Titan Xp, and the 2080Ti and maybe the 2080 have OK margins (above 10%). Lower end cards have even LESS margin. They make money purely on volume.



    Except they dont. The cost of the card is almost entirely determined by the cost of the parts and labor to assemble it. You can price the parts yourself, exept for the actual GPU die itself (but we have an idea of what those cost, as if YOU were to order something similar from Global Foundries, youd know what it was going to cost). A GTX 1060, for instance, had about a 7% margin for nVidia (the founders edition). AIB partners make even less (as they have to buy the GPUs from nVidia, instead of getting them at cost from GF or TSMC) - i remember it was EVGA that was telling Paul and Kyle (BitWit) that their average margin is sub-5%.



    How are they colluding together? Theyre selling them barely above what it costs them to make them.



    Uhh.. wrong? The 2080 performs as good as the 1080Ti or slightly better... and is the exact same price. Both are 699MSRP.



    Second-tier. The Titan is not part of the consumer product stack. And the main reason they cant sell that card for cheaper is the HBM2. If they had gone with GDDR6 they could most likely have sold that card 200$ cheaper.



    You dont have a relevant point. The 2060, btw, is 30$ CHEAPER than the previous generation product in that place in the stack. (It is NOT the replacement for the mid-range 1060, as with the removal of Titan, everything has shifted UP on product category - it is the replacement for the upper-mid-range 1070). The 1070, which it is actually the replacement for in the stack, was 30$ more expensive (380$ vs 350$) at launch.



    Only if they aren’t intelligent enough to notice that because of the removal of Titan, the previous naming scheme and how it relates to where a product sits in the stack is no longer valid.
    All tech industries, from those who make phones to CPUs have been subsidised by the public system, through grants and public research and development and appropriated by private firms and then used for profit. You won't hear or see any documents from Intel, Nvidia or AMD reporting such things because these are private industries that are unaccountable to the public. If you care enough, do your research by looking up unclassified government documents that talk about these things. Secondly cheap labour is a huge factor, with Intel using production sites in China, Philippines and Malaysia - these places are chosen not only for their cheap labour but to avoid all sorts of environmental laws and tax benefits. Lastly, every semi-credible tech source, which includes Anandtech, have reported rising prices for the mainstream. You want a top tier card in Nvidia's product stack, you are paying significantly more than last generation and that's my point. You want to compare apples to apples of different companies that's fine, but comparing a 2080 to 1080TI and given Nvidia's marketing strategy when naming their products is insane.

    You want to continue defending these poor practices, that's fine but as a consumer expect to pay more each year as a result.
    Last edited by mmoc7f933b7749; 2019-01-13 at 10:10 AM.

  14. #274
    Deleted
    Quote Originally Posted by bowchikabow View Post
    So, couple of things... I will list them in Bullet point.

    1- AMD stated that they were using an engineering sample, which means it is not the final product. Assume that IPC and frequencies WILL go up (my guess is 5 to 12% from current). And they did this using double digit reduced power levels... On an engineering sample. I am gonna guess they used R5, but will have to wait and see.

    3- AMD used an engineering sample and BEAT an all ready shipped flagship CPU. Intel doesn't have anything major set to release before 2020. Maybe it is 10nm, maybe it is 7nm. But they are, as of right now, behind the curve.
    IPC is unlikely going up - unless there is some kind of debug utility enabled in engineering sample that slows it down(unlikely). I do expect clocks to go up though and maybe power draw to go down at least on idle. I also do not think this was their best silicon. If Adored is right(and I think he is) amd beat intel core for core with lower clocks - meaning they actually surpassed them in ipc(at least in some workloads). If this demo chip was actually 4.5Ghz and we get that 4.8-5ghz one.. lets just say amd is in for a win.

    Also wth with all that gpu and build talk in here?

  15. #275
    Well IPC of that is like near haswell/skylake level.

    We are looking at 4.3 mostly...yeah it's not super fascinating but will get the job done.

    Would you wait awhile for new GPU's to launch or what would recommend after the 1060? Probably end up waiting for newer cards to get here.

    - - - Updated - - -

    [IMG][/IMG]


    Just not super Singlethreaded...but it will handle the multi-thread and work loads.

    If you asked the question yes they look like more REAL 6 and 8 cores yes.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •