|
Top end 8 core desktop part, FX-8130P so 125w TDP. Supposed to cost around $300 or so if the rumors hold up.
|
#
¿
Jul 13, 2011 06:06
|
|
|
|
| # ¿ May 15, 2024 08:48 |
|
|
Factory Factory posted:Where does the FX-8130p fit in, exactly?
|
|
#
¿
Jul 13, 2011 07:22
|
|
|
Hardocp is saying now that MS is to go with what amounts to a customized version of a BD based APU. Very much rumor mongering but would make more sense than them going with Cell for X720 or whatever. That would be a pretty big win for AMD, maybe even bigger than getting their GPU in all the next gen consoles, if true of course.
|
|
#
¿
Jul 21, 2011 11:44
|
|
|
AMD may be willing to sell them a long term liscence to the chip itself which I don't think Intel was willing to do. If so it might be made on TSMC's bulk process and not GF's custom or maybe even their bulk process which would make more sense if they're looking to control the manufacture of their next console as much as possible.Alereon posted:Keep in mind that this generation of consoles is all about cost reduction, both in terms of bill of materials but ESPECIALLY hardware development investment. I'd love to see AMD develop some custom APU with a wide GDDR5 memory bus, but I think it's more likely that any console using a Bulldozer APU will use a regular AMD Trinity APU (Bulldozer cores plus VLIW4 graphics) and a custom chipset to provide a low-cost integrated platform. PC LOAD LETTER fucked around with this message at 01:10 on Jul 23, 2011 |
|
#
¿
Jul 23, 2011 01:04
|
|
|
I was pretty optimistic about BD earlier on in the thread but I gave up waiting and give no fucks now. Microcenter had a hell of a deal on a i7 2600K + Z68 mobo bundle for less than $400 so I went that route instead. Yea the B3D guys think it was BS. They didn't give the prices but consensus in the thread was that XDR2 RAM was still too expensive. \/\/\/\/\//\/ PC LOAD LETTER fucked around with this message at 06:10 on Sep 8, 2011 |
|
#
¿
Sep 8, 2011 01:19
|
|
|
No and no clue when if ever. What leaked future release charts we have for AMD chipsets don't mention it for anything a year or 2 in the future FWIW. Seems like AMD has really fumbled BD as badly as I feared given that quote from the French site. drat shame. Not their fault TSMC is having problems too though. Seems like no one besides Intel has got their poo poo together for sub 45nm processes.
|
|
#
¿
Sep 29, 2011 22:38
|
|
|
I think AMD/Intel have got a patent agreement that covers most of that stuff. The details of it are publicaly unknown AFAIK though.
|
|
#
¿
Sep 30, 2011 03:03
|
|
|
Bloody Antlers posted:If BD stinks at launch, we could see another Phenom -> Phenom II type transition where engineers save the day by fine tuning the design and dramatically increasing performance. Yea been wondering the same thing. AMD has been very quiet on BD but they're already starting to pimp Vishera aka Piledriver. Smacks of the whole Phenom launch all over again, but then that has probably been obvious for at least a month or 2. Longinus00 posted:Bulldozer looks like it's betting not on clockspeed but concurrency and eventually offloading more and more work onto the GPU.
|
|
#
¿
Oct 3, 2011 04:15
|
|
|
Agreed posted:Again, I am still waiting for official stuff, but silence from AMD is absolutely deafening at this point given the nature of the information coming out. Fuuuuuuck. Yea, them being silent plus all the delays is what finally made me give up on BD. In theory the chip sounds like it should be pretty good while having a smaller die size but it looks like AMD hosed up their implementation. Usually when any of these tech companies have something good they "leak" info. like a sieve. When they gently caress up they get as quiet as a mouse and then start pimping the next product. AMD is being very quiet on BD but is happily hyping up Piledriver (improved BD) for next year.
|
|
#
¿
Oct 11, 2011 00:41
|
|
|
BD isn't a APU. It has no integrated GPU so no stream processors. That isn't coming until mid next year or so. Also updated BD is supposed to be around 10% faster according to the rumor mill. If that is true then there is no point waiting around for Piledriver either.
|
|
#
¿
Oct 12, 2011 21:09
|
|
|
I think that had more to do with Bobcat. Still plenty that is lol worthy in the old pre launch BD slides given what we know now though.
PC LOAD LETTER fucked around with this message at 11:38 on Oct 17, 2011 |
|
#
¿
Oct 17, 2011 11:29
|
|
|
Given what he said about BD largely panned out I don't think you can hand wave away what those ex engineers said as "disgruntled employees bitching" or exaggeration or something. As for the scheduler being the problem...I don't think anyone outside of AMD knows exactly what is wrong with BD. Most likely its a combo of several design problems and process issues.
|
|
#
¿
Oct 18, 2011 01:50
|
|
|
Agreed posted:Intel -needs- AMD. If Intel becomes an effective monopoly, nobody wins. I can only imagine the fallout from Intel getting busted up like Ma Bell, jesus. Why would you think the current administration or the next for that matter would bust them up? The last few administrations have been incredibly pro big business/mega corp and have allowed massive consolidation in many areas. They didn't even bust up the banks after they helped to trash the economy, hell they made them bigger as part of a "solution". They sure haven't broken up MS either. I'd really be surprised if they did anything if AMD essentially torpedoes itself with a bad product or something.
|
|
#
¿
Oct 20, 2011 23:59
|
|
|
trandorian posted:How would you even break up Intel? Not let the desktop and laptop cpu teams talk to each other? I'm guessing they'd break the GPU and chipset side of the business off into separate companies if they really wanted to. Not sure if even that would really make sense in the long run though given the trend to integrate everything onto a single die for power savings, to reduce the size, and cost reductions.
|
|
#
¿
Oct 21, 2011 03:06
|
|
|
Arsten posted:After reading through benchmarks and this thread, my question is "What benefits or spiffy technology are actually in Bulldozer?"  edit: actually misread that chart. It only shows performance per watt, its possible then that AMD may not be able to improve total performance much at all then but knows they can lower the power draw, which is nice but still leaves them screwed with BD. That probably means they won't be able to catch up to or beat Nehalem, which by then will be almost a 5 yr old core and likely phased out, until 2013. Excavator will probably bring them up to SB levels of performance by 2014, if they meet their performance goals of course, but by then Intel will have moved on to a better core as well. Then you have to consider the process lead Intel has as well...and yea things are looking grim for AMD. More so then ever before. Unless they can pull of some miracles in a redesign with a new chip its looking doubtful if AMD can ever catch up to Intel again. edit2: AFAIK even with all those new instructions it still loses to SB and Nehalem much less SB, at least on recompiled linux benches that were posted on RWT. There is something wrong with BD's FPU too. On paper it should be great, I used to be pretty gung ho about BD reading that stuff, but again the implementation is screwed up. That isn't to say BD doesn't get a nice speed up with recompiled apps, making it faster than PhenomII is an improvement, it just still isn't good enough. \/\/\/\/\/ edit3: of course nearly no one will recompile their apps for BD either, so its pointless talking about that seriously anyways.\/\/\/\/ PC LOAD LETTER fucked around with this message at 04:30 on Oct 24, 2011 |
|
#
¿
Oct 24, 2011 02:34
|
|
|
But then why did they blow all that die space for all that cache if it had little impact on performance and consumed so much more power? Doesn't seem to add up.
|
|
#
¿
Oct 24, 2011 13:56
|
|
|
Zhentar posted:Because it is useful for server workloads, and they only designed a single Bulldozer die. I would guess that the decision was made to conserve engineering resources.
|
|
#
¿
Oct 24, 2011 23:54
|
|
|
Zhentar posted:Removing the L3 cache may not be a big deal, but take a look at the die layout. HalloKitty posted:Those benchmarks are interesting, because they show it being competitive a fair few times. It seems like if your application is threaded enough, and can abuse some of the new features, Bulldozer is pretty reasonable (although as you said, hot/power hungry). PC LOAD LETTER fucked around with this message at 09:37 on Oct 25, 2011 |
|
#
¿
Oct 25, 2011 09:34
|
|
|
Zhentar posted:Early this year would have been way too late for that big of a change. Zhentar posted:That's at 65nm, so the 32nm BD cache should be capable of even less.
|
|
#
¿
Oct 25, 2011 16:35
|
|
|
The big thing there IMO is the support for much faster DDR3. Llano's GPU turned out to be more bandwidth limited than I thought so DDR3 2133 will probably make a real big difference.
|
|
#
¿
Oct 26, 2011 03:46
|
|
|
Factory Factory posted:Is AMD's process different, or are overclocked Bulldozer chips going to burn out faster than a fart in a frat house? Hard to find good public information but turned this up. Paul DeMone posted:AMD slide presentation Not exactly confidence inspiring. PC LOAD LETTER fucked around with this message at 17:05 on Nov 22, 2011 |
|
#
¿
Nov 22, 2011 16:56
|
|
|
Install Gentoo posted:When you get a 2 GB video card and 8 gb of system ram in a fairly cheap laptop, it'd be rather nice if the consoles we'll be using for the next 8 years have that much to work in. Star War Sex Parrot posted:might as well wait to see how the 7950 turns out. If you already have a GTX580 or 6970 it might not be worth upgrading, especially if rumors of a major refresh mid year pan out. If you're still "clunking" a long on a 4xxx or 5xxx card or if you really want to push multi high resolution monitors then its worth to upgrade IMO. For the latter application CF'd 7970's is probably sensible if expensive. So long as you keep your resolution around 1920x1080 1GB VRAM is still fine. Its not really a big problem until you higher resolutions.\/\\/\/ PC LOAD LETTER fucked around with this message at 08:14 on Dec 22, 2011 |
|
#
¿
Dec 22, 2011 07:16
|
|
|
This stuff hasn't been confirmed but lately details about AMD's new cards have been leaking all over and they don't seem totally unreasonable either. They're from the B3D thread.  You may need to spend closer to $300-350 if that chart is correct to get around 2/3 performance of the 7970. PC LOAD LETTER fucked around with this message at 23:40 on Dec 27, 2011 |
|
#
¿
Dec 27, 2011 23:38
|
|
|
Alereon posted:that's a pretty good sign the earlier rumors of POWER7 + AMD were wrong. Seems like its pretty much a given that whatever IBM/AMD/whoever do provide it'll have to be something fairly customized again to provide the performance at a reasonable cost and heat output/power consumption. Alereon posted:looks like we won't see a replacement of the FX-series Zambezi processors until the Steamroller launch in 2013, and they'll be up against Intel's Haswell processors PC LOAD LETTER fucked around with this message at 08:24 on Sep 2, 2012 |
|
#
¿
Sep 2, 2012 08:21
|
|
|
Paul MaudDib posted:The Bulldozer/Steamroller/etc architecture was just a bad idea and it looks like even AMD is giving up on fixing it. I'm reading this as capitulation and trying to figure out where to go from here. At least they're aware how hosed they are. The combo of patent issues, long development times, highly competent competition, and high production costs in making a high end x86 chip are brutally risky from a business perspective which is why no one but AMD tries (tried?) anymore to compete in that arena against Intel.
|
|
#
¿
May 8, 2014 06:01
|
|
|
Yeah the anti competitive practices by Intel certainly didn't help at all but AFAIK that was more of a way to hamstring AMD during the K7/A64 days. By the time AMD was pushing PhenomII's vs Core2 I don't think Intel was evening bothering to try that stuff anymore. e: Even during the 'good days' when their chips sold well and for good ASP's AMD was always fab constrained so I kinda hate to say it but going forward due to the increasing costs of shrinking the processes down they would've been 'doomed' to go fabless eventually. PC LOAD LETTER fucked around with this message at 08:10 on May 8, 2014 |
|
#
¿
May 8, 2014 08:07
|
|
|
They can't really do much to improve single threaded performance without a major revision (which is what Excavator is supposed to be) which takes a long time (years) to do. They keep throwing more die space at the iGPU for their APU's because for some reason they seem totally unable to increase the bandwidth to the iGPU and can't do much to improve CPU performance beyond what they're already doing so more iGPU is about the only way left for them to differentiate their product vs Intel's offerings and their own older APU's. The funny thing is if they could just feed it enough data the iGPU would actually be getting some fairly respectable performance and it would add a lot of value to their products even if they continued to do little to improve single thread performance on the CPU side. Many of their cheaper APU's in the thin n' light category are incredibly hamstrung by single channel DDR3 1333 levels of bandwidth though. Quad channel DDR3 probably isn't practical for them to do for cost reasons, same goes for on package DDR3, but maybe they could've hung some extra DDR3 or GDDR5 off of the chipset to feed their iGPU's better. They did do something like that for a while back when they were still putting the iGPU in the north bridge in the 780G or some such.
|
|
#
¿
May 9, 2014 23:54
|
|
|
SwissArmyDruid posted:Well, DDR4 is starting to ship. On package or soldered to mobo HMC/eDRAM would definitely solve their bandwidth problems for their APU's but would probably still be too expensive. You'd still have to deal with the pin/pad limit issue too.
|
|
#
¿
May 10, 2014 15:52
|
|
|
Alereon posted:The best solution for AMD APUs was to use mGDDR5, which placed ultra-fast GDDR5 memory on removable DIMMs.... Unfortunately mGDDR5 did not attract sufficient manufacturer interest and never entered the market.
|
|
#
¿
May 15, 2014 13:58
|
|
|
Hace posted:Well there is this: http://www.anandtech.com/show/7702/amd-kaveri-docs-reference-quadchannel-memory-interface-gddr5-option Alereon posted:It's not actually AMD RAM,
|
|
#
¿
May 15, 2014 22:17
|
|
|
Alereon posted:you still have to deal with Dataram for everything.
|
|
#
¿
May 16, 2014 02:12
|
|
|
keyvin posted:Can Intel invest in AMD to keep them going if they are worried about trust issues? Even when they do gently caress up colossally these mega corps are often shielded from much or even any sort of legal action and even if found guilty and fined often have their fines drastically reduced at a later date like BP did with the Deepwater Horizon oil spills or any of the banks and the robosigning scandals. Anti trust is a dead issue in this day and age and I don't know why it keeps being brought up as a serious issue anymore.
|
|
#
¿
Jan 1, 2015 05:04
|
|
|
I didn't know that. Should've suspected that though.
|
|
#
¿
Jan 1, 2015 07:27
|
|
|
They're gonna put Carrizo in some slim and SFF desktops so they haven't totally abandoned the desktop market but yea no socketed version and performance still won't be on par with Haswell much less Skylake. Not good at all. They better hope they do reasonably well in the cheap laptop market. Either way it'll be a looong way until Zen and their new x86 arch. comes out. Early-mid 2016 is the expected launch date I think for their new ARM arch. Haven't seen much info. about the new high end x86 arch. they're developing. They should both be platform compatible so if AMD isn't fudging things then you can plug either of them into the same motherboard so long as the sockets match up.
|
|
#
¿
Jan 11, 2015 14:04
|
|
|
G34 has all those pins for extra memory slots + allowing CPU's to communicate, its not needed for HBM or any other on package memory really. Socket size isn't the limitation for HBM. Its price that could (probably will) kill the idea for low cost APU's which is the way AMD has to price them in order to sell them. Slapping HBM on their APU's might allow them to get mid-tier-ish GPU performance but also means they'll have to price them lots higher just to break even. WAG on my part here but $2-300 would probably be the price range of a 'high end' APU with a 1-2GB HBM cache on package via interposer. Even tied to Excavator CPU's the performance vs price wouldn't be bad but most enthusiasts, and those are the ones who'd be interested, probably wouldn't bother with it for that price. A low end Intel chip and a mid range dGPU would probably be better value over all even if it ends up costing a bit more. It sucks but I think they're stuck being bandwidth limited with their APU's for a long time.
|
|
#
¿
Feb 14, 2015 10:22
|
|
|
Supposedly its not so much the memory costs that get you with HBM. Its the cost of the interposer, which is effectively a huge die IC fabbed on a 'coarser' process, plus the cost of assembly and testing. The interposer has to be gigantic since it sits under the die, Fiji has a rumored 550mm2 die while Carrizo has a ~250mm2 die size, and must be still big enough to allow the HBM memory packages to be mounted to it too. Smaller multi die interposers to bring down costs are something that is being worked on but I don't think its working outside of a lab yet. IIRC HBM2 is about giving you more bandwidth and/or RAM capacity and not necessarily bringing down costs. Depending on how they configure the GPU (ie. going all out for more bandwidth or RAM capacity) you might see costs go up or stay the same.
|
|
#
¿
Feb 14, 2015 11:51
|
|
|
The HBM packages themselves are tiny. About the size of a small pill, so size isn't an issue for putting them on a FM3 package. Stolen from this thread which BTW has tons of great info. and discussion on the subject about HBM and Fiji in general. It was other types of RAM that would've had possible packaging issues with FM3 due to size. I suppose they could've just left the pin out the same size and extended the package out to one side and plopped the RAM there if they really wanted to use something off the shelf and not do a custom RAM package. There is no indication that they ever wanted to go that route though. Zen we have hardly any information, branding or otherwise, at all on so only the most wild of WAG's are possible about it at this point. I would think if they can get near-Skylake performance at similar TDP they'll be thrilled. Skylake-ish performance with a significantly higher (10-25w) TDP is probably more realistic given the process and R&D disparity. I hope they can pull that off. If they do and they sell it for a bit less, like historically they did normally, it could be a successful chip. PC LOAD LETTER fucked around with this message at 16:06 on Feb 14, 2015 |
|
#
¿
Feb 14, 2015 16:03
|
|
|
Yea it'd probably fit. Even at 500mm2 the die would be still quite smaller than the package. Would they actually do that? I dunno. Kinda doubt it. That'd be an expensive die that they couldn't sell for a low price.
|
|
#
¿
Feb 15, 2015 01:40
|
|
|
That will probably change over time though. The Chinese market is too potentially big for them not to tune their game line up and adjust pricing to accommodate.
|
|
#
¿
Mar 15, 2015 04:49
|
|
|
|
| # ¿ May 15, 2024 08:48 |
|
|
Nintendo Kid posted:No not really. They haven't bothered to really do it for places like Brazil and India, why do it for China?
|
|
#
¿
Mar 15, 2015 23:15
|
|