AMD held a press briefing today on their upcoming 8000M graphics chips, which they are calling the "second generation GCN architecture" parts. We’ll have more on that in a moment, but while we were expecting (dreading) a rebranding prior to the call, it appears we are at least partially mistaken; there will be at least one completely new GPU with 8000M. (If you want additional background material, you can see the previous generation high-end 7000M announcement from April 2012 for reference.)

I’m not going to get too far into the marketing aspects, as we’ve heard all of this information before: AMD has improved Enduro Technology, they’re continuing to improve their drivers, and APP Acceleration has a few more applications. There have been a few major titles released in the past couple of months with AMD Gaming Evolved branding (Far Cry 3 is arguably the most notable of the offerings, with Hitman: Absolution and Sleeping Dogs also scoring well amongst critics and users), and Bioshock Infinite is at least one future release that I'm looking forward to playing.

Cutting straight to the chase, at this point AMD has released limited information on the core specifications for some of their 8000M GPUs, but they coyly note that at least one more GPU announcement will be forthcoming in Q2 2013 (8900M by all appearances). Today is a soft launch of high level details, with more architectural information and product details scheduled for January 7, 2013 at CES. AMD did not share any codenames for the newly announced mobile GPUs, if you’re wondering, other than the overall family name of “Solar” for the mobile chips (replacing the outgoing “London” series), but we do know from other sources that the 384 core part is codenamed "Mars" while the larger 640 core part is codenamed "Neptune". Here are the details we have right now:

AMD Radeon HD 8500M, 8600M, 8700M, and 8800M
HD 8500M
HD 8600M
HD 8700M
HD 8800M
Stream Processors 384 384 384 640
Engine Clock 650MHz 775MHz 650-850MHz 650-700MHz
Memory Clock 2.0GHz/4.5GHz 2.0GHz/4.5GHz 2.0GHz/4.5GHz 4.5GHz
FP32 GFLOPS 537 633 537-691 992
FP64 GFLOPS 33 39 33-42 62

Obviously there are a lot of missing pieces right now, but what we immediately notice is that the core count on the 8500M/8600M/8700M means that we’re definitely looking at a new GPU. The only other time we’ve seen AMD do 384 cores is with Trinity, but that’s a VLIW4 architecture so we’re not seeing that again. Given the currently shipping Southern Islands chips (“London” on the mobile side) have 640 cores max for Cape Verde, 1280 max for Pitcairn, and up to 2048 for Tahiti, AMD has likely created a fourth SI derivative that drops down to two CU arrays, each with three CUs. (You can read more about the GCN/SI architecture in our earlier GPU coverage.) Performance is something of a wildcard with the new 384 core parts, and the choice of DDR3/GDDR5 memory will also influence the final result. We'll find out in the coming months how the 8500/8600/8700M stack up to NVIDIA's midrange "GT" offerings, which interestingly are also using 384 cores.

Also worth a quick note is that AMD is not discussing TDPs at this point in time—which is common practice for both AMD and NVIDIA. We expect the new "Mars" parts to be more power efficient than the outgoing Thames/Turks cores, thanks to the shrink to a 28nm process. However, AMD and NVIDIA typically stick to common power targets for laptops that are dictated by their OEM partners, which often means they'll play with clock speeds in order to hit a specific TDP. That's why all of the clock speeds listed in the above table have a qualifying "up to" prefix (which I omitted).

The final announced card is the one where we appear to have more of a rebrand/optimization of a previous generation chip. 8800M has the same 640 core count as Cape Verde/7800M, only with modified clocks this time. The earlier 7800M chips could clock up as high as 800MHz, so maximum core clock is actually down a bit, but they only ran the memory at up to 1GHz (4GHz effective) GDDR5. If AMD determined memory bandwidth was more important for that particular GPU than shader performance, the new 8800M would make sense. Also note that AMD isn’t including the boost clock speeds into the above chart; under the right circumstances, all of the new chips can run at higher clocks than the reference clock.

Radeon 7800M Left, Radeon 8800M Right

AMD isn’t calling the 8800M a rebrand, but we’re looking at the same core counts as Cape Verde and the same 28nm process technology, so we wouldn’t expect a substantial change in performance. There’s also the above chip shot as a point of reference. If the 8800M is substantially different from Cape Verde then the above images provided in AMD’s slides must be incorrect, as the new and old chips look the same. Minor tweaks to power use, caching, or other elements notwithstanding, we’re probably dealing with a die respin at most. But, there’s nothing inherently wrong with rebranding—AMD and NVIDIA have both been doing it for some time now. Don’t expect every “upgraded” GPU to be better; a 7400M isn’t faster than a 6700M, and likewise we expect 7700M and 7800M to be faster options than the 384 core 8500M/8600M/8700M and competitive with 8800M. Here’s a quick recap of the same core specs as above for the current 7700M/7800M parts:

AMD Radeon HD 7700M/7800M Specifications
HD 7730M
HD 7750M
HD 7770M
HD 7850M
HD 7870M
Stream Processors 512 512 512 640 640
Engine Clock 575-675MHz 575MHz 675MHz 675MHz 800MHz
Memory Clock 1.8GHz 4.0GHz 4.0GHz 4.0GHz 4.0GHz
FP32 GFLOPS 589-691 589 691 864 1024
FP64 GFLOPS 36.8-43.2 36.8 43.2 54 64

I’ll refrain from commenting too much more about performance of an unreleased part, but AMD indicated their 8870M should be substantially faster than NVIDIA’s current GT 650M GDDR5 (which isn’t too surprising considering clocks and core counts), and the 8770M should likewise be a healthy 20%+ bump in performance relative to the 7670M. I’d rather see comparisons with GTX 670MX and HD 7770M, respectively, but I suspect those wouldn’t be quite as impressive. Anyway, you can see AMD’s comparison charts in the complete slide deck gallery below. Availability of the new GPUs is slated for Q1 2013.



View All Comments

  • JarredWalton - Tuesday, December 18, 2012 - link

    I don't care if stuff sells like hotcakes if it doesn't make sense. There are a bunch of business laptops that ship with GPUs like the HD 7470M, but there's not really a good reason to put such a GPU into a business laptop. Either you need more performance than HD 4000, and in that case you need HD 7600M or above, or you don't need it at all. The reason it sells is because the OEMs push it, not because it makes sense.

    Quit being obtuse and suggesting that I've ever stated Mars will be worse than Turks or the other 40nm parts. It's not a question of whether it will be better, but rather how much better. 28nm parts will be faster than the existing 40nm parts, they'll likely use a bit less power under load for the improved performance, and they're going to cost a lot less for AMD to manufacture. But they'll cost about the same as the current parts, and they will still be a large step down from the true midrange parts. OEMs will push them in their crappy laptops that sell thousands of units but are built of plastic and fall apart in a year or two, and in 90% of those laptops they won't even matter because the owners won't play games.

    You're hyperbole is so far out of line that I have to think that this is more of a personal vendetta than a reasonable discussion. "Exponentially worse"? Oh, wait: it's the Internet. You should be happy I even took the time to respond to your attack laden message, as really you're just spouting off with little useful data to back it up. You're extrapolating based on what you think will happen, I'm saying "we'll see when product ships." If you want a good gaming GPU, the new Mars chips are about as low as you can go without having to dial way back on details. Personally, I'd rather go with the already shipping Cape Verde/Chelsea GPUs.

    Based on what I've discussed in the above posts, we'll see a 10-20% increase in performance, relative to Turks, and that's good but not earthshattering. AMD on slide 25 shows a performance increase of roughly 30% going from 7670M to 8770M. If they're the same price, that's great, which is what slide 28 might suggest; but slide 28 also suggests that there is no currently slated replacement for HD 7800M or 7900M. 7900M I can understand, but 8800M is 7800M in every way we can find, probably a bit slower thanks to the lower clocks, so I don't know what AMD is thinking there. Will we have 8970M as the top-tier part and 8950M as the 7870M replacement? Maybe 8970M-X2 for the top with 8970M as the next step down?

    You're reading this marketing chart as though it's the bible and is not to be questioned; my job as an analyst is in fat to look at stuff like this and question it. So I look at the charts and say, "Gee, 8800M is a rebrand of 7800M, which AMD refuses to admit; AMD pulled this launch in three weeks (from the Jan 7 official launch) and gave us a bunch of fluff. Why?" Well, MrDude says it's because the new low-end GPUs will be awesome, but my opinion -- and the sentiment of many other readers -- is that the low-end GPUs are only interesting for OEMs. At the higher end segments where having a dGPU really matters, we're getting a lateral rebrand at best. The OEMs win, AMD wins, and the sheep buying entry level to midrange laptops lose. Congratulations!

    Never mind the drivers, which are still in need of work (though thankfully getting better), or the fact that NVIDIA has offered performance that is around 40% better than Turks with lower power requirements for over six months. Yes, GT 640M has been kicking the crap out of anything using HD 7670M and below since it launched. Meanwhile, 7730M is in one laptop I can find right now (Dell Inspiron 15R Special Edition) and the 7750M is also in one laptop (HP ENVY 15), while 7770M is MIA. And naturally, Enduro hasn't really worked properly/adequately until the Catalyst 9.01 beta. So AMD now has an SI chip coming in Q1 that will hopefully be able to match GK107 on performance and power, and maybe even at a lower price. But until we can buy the laptops, we can't really say for sure.
  • mrdude - Thursday, December 20, 2012 - link

    A crappy mobile GPU that isn't HD4000 on a ULV machine also means no stuttering, which any Intel ULV will do, and that's without question. The only question is if you're going to stutter due to TDP constraints or will it overheat before it gets there.

    And practically nothing of what you've wrote in the comment is implied or referenced in the article. And I'm not being obtuse, I'm simply pointing out your ignorance here. I'm not the one who claimed that anything 7670m and lower isn't worth mentioning. You did. You chose only to focus on the higher end parts and said nothing regarding the performance bump, which we both agree we'll inevitably see. The price points, too, were completely ignored. Because, as you said, only high end matters.

    And lastly, you're confusing me with someone with a sentimental attachment to a corporation which won't see the light of day in 2014. I wouldn't take an AMD GPU in a laptop if given to me for free due to the driver issues, but if labeling me an AMD fanboy makes you feel better for a horrible article and unbelievably ignorant comments then keep up the good work.
  • JarredWalton - Monday, December 24, 2012 - link

    I said "Performance is something of a wild card depending on the use of DDR3/GDDR5" -- and I could add 128-bit or 64-bit interfaces (wouldn't be surprised to see some 64-bit stuff come out still, though I hope that's not the case), and the "up to xxxMHz" clauses always make me leery. I expect an improvement on the 7670M and lower GPUs ranging from about 0% to as much as 30%, depending on the title. But as I've been trying to make clear, it depends on the comparison you make.

    With 8570M be faster than 7670M? It depends on final clocks, but probably a difference of less than 10%. I draw the line of "useful improvement" from a hardware update at around 20%, because a 10% improvement just doesn't matter enough in most cases. So when AMD shows 8690M vs. 7590M and gets a 30% average improvement, I have to point out the fact that they're skewing the numbers by intentionally using a higher ranking 8000M part vs. a lower ranking 7000M part, which makes me wonder what else we're not seeing.

    That is all.
  • JarredWalton - Monday, December 24, 2012 - link

    Also note what I say in the rest of the paragraph, which is why we're really waiting for performance details on real, shipping hardware: "We'll find out in the coming months how the 8500/8600/8700M stack up to NVIDIA's midrange "GT" offerings, which interestingly are also using 384 cores."

    You can bellyache all you want about me not talking about how AMD is improving on AMD, but that's not the question. If HD 8500/8600/8700M can't outperform NVIDIA's (now six months old) GT 640M/650M/GTX 660M, then this update is definitely too little, too late. And yet again, the slides from AMD compare the 8870M with the GT 650M, which is a completely lopsided match.

    So, I'm not ignorant (as in, there's not something useful in this conversation that I don't know), but you *are* still being obtuse (slow to understand). If you weren't being a typical Internet troll and thumping your chest, we could have agreed pages ago and not worried about the trivial differences of opinion. You have your opinion that this is a major update; I understand that opinion and I disagree that it's a big deal -- at least from the perspective of the users. It's great for AMD (cheaper to make parts at similar prices means better profits -- something AMD desperately needs), but as I have repeatedly stated, if you want a GPU for something, you likely want more than HD 7600M/8700M hardware. I can "get by" with GT 650M, but it's not an awesome experience for the times when I need a GPU (gaming); it's at best tolerable.
  • just4U - Friday, December 28, 2012 - link

    I am glad you took the time to reply so much on this article Jarred. While going hands on in the forum can have it's pitfalls it also helps. Thanks! Reply
  • CeriseCogburn - Wednesday, December 19, 2012 - link

    Thanks for soliciting the truth from the author:

    " Never mind the drivers, which are still in need of work (though thankfully getting better), or the fact that NVIDIA has offered performance that is around 40% better than Turks with lower power requirements for over six months. Yes, GT 640M has been kicking the crap out of anything using HD 7670M and below since it launched. "

    LOL - can you say amd epic fail, over and over ?

    Here, let me show you how nice the author was to you, and your fanatic fanboy stance:

    " Today is a soft launch of high level details, with more architectural information and product details scheduled for January 7, 2013 at CES"

    WUT!?? A "soft launch" ? No, this is a complete PAPER LAUNCH 100%.

    See how that goes, with amd, they can announce and after endless years of screaming paper launch toward nVidia, it is casually announced as a "soft launch". Not a single freaking chip in sight, but it's all soft and lovely when amd does it...

    Why weren't you praising the author for covering amd butt ?

    " Granted, these are all preliminary results and we won't have hardware for a month or more I'd guess. We don't know power or other aspects of performance that are important either. "

    Hey wonderful ! Not a chip in sight, but it's a "soft launch", landing so happily in the breast of amd fanboys...

    You complain too much amd fanboy.
  • JarredWalton - Thursday, December 20, 2012 - link

    Note that "soft launch" and "paper launch" are nearly the same in my book, and don't forget that we've pretty much never had a hard launch of mobile GPUs. You don't buy mobile GPUs separate from a laptop, so when the hardware is launched and available (e.g. the new 8000M chips), that's when notebook manufacturers are going to start putting chips into laptops. It takes a few months of testing and validation before they start releasing those to the public, and thus we have a "soft launch". Reply
  • lmcd - Monday, December 17, 2012 - link

    The whole point of the lower-end parts is to pair with Kaveri for Hybrid XF, in which case you want it to be on the same architecture (VLIW5 with VLIW4 not being a major issue). Meanwhile, Intel laptops will use high-end ones which *are* improvements or the same (such as 88xx or 89xx) or none at all (in OEM fashion).

    So who cares.
  • iMacmatician - Tuesday, December 18, 2012 - link

    "Also I can't wait to see 8700M parts clocked as low as 8500M, as the graph suggests it will be possible. Disgusting."

    Well, the 8700Ms may still have a performance lead, as (according to they have a 128-bit bus while the 8500M/8600Ms have a 64-bit bus.
  • CeriseCogburn - Wednesday, December 19, 2012 - link

    Dear mr dude, to impart this is important. From your whining link:

    " The 8000M series doesn't appear to boast any unique architectural refinements, though. AMD told us that, while these parts do feature new silicon, they're still based on the original iteration of the GCN architecture, just like the desktop Radeon HD 7000 series. "

    LOL - Can you say rebrand ?

    " Functionality like DirectX 11.1 support, PCI Express 3.0 connectivity, and AMD's Enduro technology are all included, but that's also true of the GCN-based 7000M-series parts. "

    Say it again, amd fan.

    I'll give amd this much - they spin and lie and spew so much CRAP - they have their little reporter minions spinning circles around themselves.... and thus immense confusion, many hopeful and decieved once again amd fanboys start their little hearth and happy fires burning - and the direct readership is left spinning in circles.... not certain what just occurred....

    This is of course amd's plan, and it works wonderfully.... so of course sanity must be injected...

    CUT DOWN REBRANDED GPU ( look for that on the laptop )

    :-0) always gald to help.

Log in

Don't have an account? Sign up now