|
Factory Factory posted:Hey, you never know, Bulldozer could completely change the way we think about being a complacent, slightly depressing runner up! Me too, I do love my Intel CPUs, but competition is very good and we all know AMD is capable of great things. They crushed Intel for a good two years with the Athlon 64 running circles around the Pentium 4, maybe they can do it again!
|
#
?
Sep 23, 2011 17:22
|
|
|
|
| # ? May 8, 2024 21:04 |
|
|
They've demonstrated that it's a good overclocker. I imagine that's going to be their push in the desktop market. Intel seems better positioned (due to having dump-trucks full of hundred dollar bills) to bring their prices down if it means staying competitive, though. At the proposed price points, AMD is going to have trouble, especially since Intel's going to be offering the 2700K at the 2600K's price point (which is probably just going to be a binning process with the current 2600Ks, but still, it'll push prices on the 2600K and 2500K down by design to make it that much harder for AMD). AMD has talked about "easily getting 5ghz on air" as part of their press release regarding their record-holding 8GHz+ clock on a Bulldozer chip; the only problem is that their claimed clock for clock improvement, unless it's intentional misdirection (why would it be, let their stock take a dive while Intel pushes forward with high-powered architecture and even greater power efficiency for the markets where the real money is?) means that if every single Bulldozer buyer gets 5GHz no problem, it's just a big number. Assuming 35% improvement over current AMD CPUs, that's still lagging behind Intel's performance, and the clock advantage means that even if most Intel CPUs aren't likely to hit 5.0GHz right now, most of them will do 4.4-4.5GHz, so even among enthusiasts they've got an advantage in performance at the same price point as AMD's Bulldozer chips. I guess it remains to be seen what the more-than-just-hyperthreading, less-than-additional-cores extra parts on the Bulldozer chips will do for performance, but they better all overclock like crazy if AMD's got a shot at regaining some performance competitiveness this generation. And bring the server parts down, Intel runs the market there, how will AMD compete apart from cost savings?
|
|
#
?
Sep 23, 2011 18:17
|
|
|
movax posted:good time to have Intel stock.  There is literally no good time to have Intel stock. Trust me, this poo poo never moves. Remember when we announced the best year in company history last year? Yeah that stock bump of a buck lasted a week maybe.
|
|
#
?
Sep 23, 2011 18:20
|
|
|
Agreed posted:They've demonstrated that it's a good overclocker. I imagine that's going to be their push in the desktop market. Intel seems better positioned (due to having dump-trucks full of hundred dollar bills) to bring their prices down if it means staying competitive, though. At the proposed price points, AMD is going to have trouble, especially since Intel's going to be offering the 2700K at the 2600K's price point (which is probably just going to be a binning process with the current 2600Ks, but still, it'll push prices on the 2600K and 2500K down by design to make it that much harder for AMD). Llano is taking a monstrous 1.5+ volts to hit a measly 3.8GHz with 32nm which is terrible consider the same architecture hit 4GHz at less voltages with 45nm. My hunch is all these BD delays were due to poor 32nm yields, and since BD was designed to hit 4GHz+ from the ground up I don't really think BD will be a winner on performance/watt front if the voltages are still sky high.
|
|
#
?
Sep 23, 2011 18:57
|
|
|
Henrik Zetterberg posted:
The stock goes up to maybe $24+ after record earnings and generous dividends are announced. Then a couple of weeks later it goes back down like nothing happened.
|
|
#
?
Sep 23, 2011 19:36
|
|
|
Asus demos PCIe Gen 3.0 with an Ivy Bridge CPU. They're stepping around the lack of generally available PCIe 3.0 silicon by using a PCI Express switch, the PEX8747 from PLX and modifying the upstream port link width from a x1 Gen 3 to a x1 Gen 2 (1GB/s vs. 500MB/s). The PLX 8700 series is Gen 3 compliant on all ports. Extra sperg facts: The 8747 can have up to 5 ports in various configurations that could simulate what an Ivy Bridge system would do. x16 upstream to 2 x16 endpoints, x16 upstream to 2 x8 endpoints + 1 x16 endpoint, or a x16 upstream to 4 x8 endpoints. We'll probably not see this on motherboards though; for reference the 8648 (generation 2 switch big boy) is $117 per unit right now. Obvious difference in performance, but x16 Gen 2.0 still isn't fully stressing GPUs, as HardOCP has shown in their tests. Where this would pay off I think though, is in being able to route less lanes to a slot for the same effective bandwidth, or what Intel is planning on doing with Panther Point, which is some chipsets allowing the CPU to split its lanes to more than just 1x16 or 2x8; you could get 1x8 + 2x4, which could be nice for RAID controllers.
|
|
#
?
Oct 5, 2011 06:02
|
|
|
I hate replying to myself but I have a bump for a legit question that I seem to have forgotten the answer too and can't find: what's the interconnect between the Sandy Bridge GPU and the system as a whole? I know the GPU can share L3 cache with the CPU, so is it a proprietary interconnect that still exposes it in PCI space, or is it an internal PCI Express link to the RC in the CPU? (if you're running one right now, go into Device Manager -> View devices by connection and post what it shows up under)
|
|
#
?
Oct 6, 2011 15:58
|
|
|
I thought it's connected to that new magical ring bus, and the CPU pretends it being on PCIe with trickery in its PCIe core?
|
|
#
?
Oct 6, 2011 16:15
|
|
|
Combat Pretzel posted:I thought it's connected to that new magical ring bus, and the CPU pretends it being on PCIe with trickery in its PCIe core? That's what I was thinking as well, just wanted to make sure. The latter part is definitely true, it's exposed as a standard PCIe device for software's sake.
|
|
#
?
Oct 6, 2011 16:19
|
|
|
If you somehow got an Ivy Bridge that fell off the back of a truck, Asus has already released beta BIOS updates that will support it.  Also, Intel RST may need an option ROM update to supported >=2.2TB disks, so check for a BIOS update. It's fine in AHCI mode I guess, but running the PCH in RAID mode, older ones won't support 3TB drives.
|
|
#
?
Oct 9, 2011 18:36
|
|
|
Not sure, possibly I'm going to skip it. In theory, power savings could possibly make up for it relatively fast, seeing how my box runs 24/7. Some changes listed in that Anandtech article earlier could prove advantageous for the hyperthreaded variants. And some math improvements (that divider) could help in my photo, video and rendering stuff. We'll see.
|
|
#
?
Oct 9, 2011 19:45
|
|
|
Combat Pretzel posted:Not sure, possibly I'm going to skip it. In theory, power savings could possibly make up for it relatively fast, seeing how my box runs 24/7. Some changes listed in that Anandtech article earlier could prove advantageous for the hyperthreaded variants. And some math improvements (that divider) could help in my photo, video and rendering stuff. We'll see. I definitely don't think it's worth upgrading too from Sandy Bridge, agreed. Possibly not even Nehalem, seeing as that generation is still more than sufficient for games (hell, even the Q6600 OC'd is acceptable). But if someone's in the market for a brand-new machine in 2012 coming from an older AMD or a C2D system, it'll definitely be great for them.
|
|
#
?
Oct 10, 2011 05:40
|
|
|
Yeah, this is an extraordinarily impressive power efficiency improvement. I'm not knocked out by the clock for clock - I remember the 20% figure, obviously they meant power:performance:clock rather than just 20% more, eat it Sandy Bridge - but I am impressed by it nonetheless. Seems like a very well put together processor and the move to 22nm and a superior transistor construction had to happen sooner or later, what better time to get it up and running than when your nearest competitor is dead in the water, I guess. Their current laptop lineup is already extremely power efficient, I imagine Ivy Bridge is going to kick some real rear end in that market. Most importantly it remains very clear that Intel knows what the hell they're doing and continue tick, tock, tick, tock while everything that could go wrong is going wrong for AMD. I don't view that as a good thing, really, but they sure do know how to keep to a schedule. drat.
|
|
#
?
Oct 10, 2011 05:49
|
|
|
movax posted:I definitely don't think it's worth upgrading too from Sandy Bridge, agreed. Possibly not even Nehalem, seeing as that generation is still more than sufficient for games (hell, even the Q6600 OC'd is acceptable). But if someone's in the market for a brand-new machine in 2012 coming from an older AMD or a C2D system, it'll definitely be great for them.
|
|
#
?
Oct 10, 2011 05:51
|
|
|
Alereon posted:What could be interesting is the potential for drop-in upgrades of cheaper i3 machines to Ivy Bridge i5s, that could provide a very meaningful upgrade for not too much money. Doubling the cores, adding Turbo Boost, and vastly improving the graphics should help a lot. Definitely, my only concern would be if the motherboards those i3s would usually be found in would eventually receive BIOS updates to support the new CPUs or not.
|
|
#
?
Oct 10, 2011 05:53
|
|
|
I've been really impressed by the video encoding capabilities of the Sandy Bridge (I have some background doing hardware for h.264/vc1 encoders). I don't own one yet but from messing around with ones we have at work that quick sync video is the bee's knees. Definitely planning on getting an Ivy Bridge in my next upgrade cycle.
|
|
#
?
Oct 10, 2011 05:55
|
|
|
QuickSync is darn fast, but the encoding quality is still pretty bad compared to a good software encoder like x264. I hear that's an area they'll be improving in Ivy Bridge though, and hopefully there will be better encoding tools available by then.
|
|
#
?
Oct 10, 2011 06:18
|
|
|
Bit-Tech has an Ivy Bridge architecture article up. It doesn't look like it has much new information, but it's got a pretty easy-to-understand explanation of how 3D and tri-gate transistors work.
|
|
#
?
Oct 12, 2011 04:06
|
|
|
priznat posted:I've been really impressed by the video encoding capabilities of the Sandy Bridge (I have some background doing hardware for h.264/vc1 encoders).  ) even faster... kind of makes QS redundant aside from one-time library rips and that sort of thing.
|
|
#
?
Oct 12, 2011 04:33
|
|
|
Yeah it'll definitely need some quality software developed for it, but the offload capabilities in the CPU is a pretty nifty thing. It's just kind of mind blowing considering what a decent hardware encoder was like 5-6 years ago (just like all computer related stuff, I know  ). Decent motion estimation alone took up a high end and bloody expensive FPGA and all that.
|
|
#
?
Oct 12, 2011 04:44
|
|
|
priznat posted:Yeah it'll definitely need some quality software developed for it, but the offload capabilities in the CPU is a pretty nifty thing. Well, if it can be done on an FPGA then you could also do it on a regular ASIC. I'm pretty sure you could write a smoking encoder targeting the PS3s cell chip if you wanted. It's economics that dictate what route you take. VVV Why would you want a 3TB boot drive anyway? It's much better to make a small fast SSD the boot disk. Longinus00 fucked around with this message at 06:26 on Oct 12, 2011 |
|
#
?
Oct 12, 2011 06:19
|
|
|
I really hate how I picked the last major hardware generation before EFI (i5 750, P7P55D). All you guys with your mice in EFIBIOS and 3TB boot drives  .
|
|
#
?
Oct 12, 2011 06:24
|
|
|
Longinus00 posted:Well, if it can be done on an FPGA then you could also do it on a regular ASIC. I'm pretty sure you could write a smoking encoder targeting the PS3s cell chip if you wanted. It's economics that dictate what route you take. Well that's for sure, this was for a low volume/high margin broadcast video encoder so FPGAs made more sense at the time than a full blown ASIC development. It's just neat that Intel is pushing so many features into the CPU is all.
|
|
#
?
Oct 12, 2011 06:33
|
|
|
priznat posted:Well that's for sure, this was for a low volume/high margin broadcast video encoder so FPGAs made more sense at the time than a full blown ASIC development. 'sup FPGA buddy  I posted this in another thread somewhere, but I'm honestly not a fan of all the hardware acceleration coming out for video encoding and decoding, because the silicon is being flung at us developers rapidly without time or resources to properly implement support. Even things like DXVA were kind of broken at first, but at least that's now relatively mature in MPC-HC. I mean just off-hand, for decoding, I've got CUDA, Ions, Broadcom's CrystalHD, etc. Maximum compatibility and the least amount of issues shows up when you just let the CPU do the brunt of the work, and well, that's what the raw horsepower of Sandy Bridge (and past generations) is for. My E6600 could do 1080p with the rare dropped frame; my 2600K will do the same whilst pegging maybe 15% in Task Manager. For encoding, there are custom commercial encoders that have CUDA ports, or leverage Quick Sync if available. But for something open-source like x264, that's developed by talented people in their spare time, and the priorities are somewhat different. If all you need is phone quality rips for your iDevice or something, then hell yeah, Quick Sync might be the way to go. For any serious archival/etc needs though, you're probably using a fairly involved encoding chain that uses x264 as the H.264 encoder. What is neat is Ivy Bridge's (IIRC) new DRNG; there's an interesting overview on it on IEEE Spectrum covering how to use pure digital logic to generate random numbers.
|
|
#
?
Oct 12, 2011 06:46
|
|
|
Haha, it's true, video offload stuff is crammed into pretty much everything these days. I had picked up one of those Broadcom crystal HD things in order to put it in an apple tv v1 but nothing came of that (at least it was cheap). I miss working on AVC stuff, that was fun! Except for CABAC, that poo poo will give you such a headache
|
|
#
?
Oct 12, 2011 07:21
|
|
|
To cross-post a bit from the AMD thread, Bulldozer is out and it is in no way competitive with Sandy Bridge. Even multi-threaded performance is terrible, which was the one area it was supposed to beat Intel.
|
|
#
?
Oct 12, 2011 07:29
|
|
|
priznat posted:Haha, it's true, video offload stuff is crammed into pretty much everything these days. I had picked up one of those Broadcom crystal HD things in order to put it in an apple tv v1 but nothing came of that (at least it was cheap). Yeah, CABAC is a bitch, and the whole flexible macroblock ordering isn't really very fun either. I guess I'm a masochist though, because I can't get enough of video compression/related technologies. Alereon: hey, at least everyone who bought a 2500K/2600K can rest easy knowing their e-peens are safe for the moment!
|
|
#
?
Oct 12, 2011 07:34
|
|
|
And all those hardware encode solutions don't work for poo poo for real time applications. That's a pretty terrible launch for AMD. Were they expecting much higher clockrates? I mean, the 2500k can be OCed by a good 33% on air as well. drat.
|
|
#
?
Oct 12, 2011 15:17
|
|
|
Fuzzy Mammal posted:And all those hardware encode solutions don't work for poo poo for real time applications. Remember too that AMD is fabless, whereas Intel not only own their own fabs, but has been a huge DRAM manufacturer for a long while as well. They have an impressive list of people with "PhD" after their name that developed many of the techniques used in semiconductor manufacturing today. I don't doubt that AMD has some excellent process engineers, but they depend on TSMC and friends' process offerings to do their products justice.
|
|
#
?
Oct 12, 2011 16:21
|
|
|
Fuzzy Mammal posted:And all those hardware encode solutions don't work for poo poo for real time applications.
|
|
#
?
Oct 12, 2011 16:28
|
|
|
They were publicly saying 50% clock for clock. Then publicly lowered expectations to 35% clock for clock. Which is obviously a crock of poo poo for the majority of applications, and a patch isn't going to fix that. Edit: Oh, sorry, I get what you mean. Clock overall, not clock for clock. 
Agreed fucked around with this message at 16:33 on Oct 12, 2011 |
|
#
?
Oct 12, 2011 16:29
|
|
|
Alereon posted:To cross-post a bit from the AMD thread, Bulldozer is out and it is in no way competitive with Sandy Bridge. Even multi-threaded performance is terrible, which was the one area it was supposed to beat Intel. Was whoever is currently in charge at AMD asleep 2000-2006?
|
|
#
?
Oct 12, 2011 17:40
|
|
|
SRQ posted:So AMD and Intel completely reversed roles in ten years, with AMD shipping crappy power-hungry high clock chips, and Intel shipping efficient faster lower clock chips.
|
|
#
?
Oct 12, 2011 17:56
|
|
|
Alereon posted:AMD recently fired their CEO, and the new CEO fired the General Manager of the Products Group, now I guess we know why (though a lack of mobile strategy was the general reason cited for firing the CEO). Those guys got a tremendous amount of sympathy from all the enthusiasts and pundits at the time. I wonder how much they would have gotten if they were fired after these reviews.
|
|
#
?
Oct 12, 2011 18:01
|
|
|
Delicious new slides: Looks like P1 (Performance) still gets HT/8M cache, mainstream is stuck without those. Max TDP of 77W!  Pretty cut and dry, first digit gets bumped up to 3xxx. Pentium does...whatever Pentiums do.  quote:As for backwards and forwards compatibility, well, things are a little bit complicated. All Sandy Bridge CPU's are according to Intel, guaranteed to work on all the new 7-series chipsets, so if you want a new motherboard for your old CPU, then you don't need to worry. However, if you want to use a current motherboard with an Ivy Bridge CPU, then things get a little trickier. For starters, the Q67, Q65 and B65 chipsets will not support Ivy Bridge and if we're correct, this has to do with lack of firmware space. This might once again only be limited to Intel's motherboards though, as the company didn't put large enough flash chips on its motherboards. I guess some boards that are only sporting 2MB SPI ROMs are SOL. 4MB chips are still good, as ASUS already has a beta 2001 BIOS out for their boards to support Ivy Bridge. Looks like everyone is still waiting on a final version of Management Engine (:jerkoff:) before they can get their poo poo straightened.
|
|
#
?
Oct 18, 2011 05:16
|
|
|
The Something Awful Forums > Discussion > Serious Hardware / Software Crap > Intel Roadmap Discussion: Ivy Bridge Roulette Spring 2012. Are they still going to stick it to the K proc buyers and leave out Virtualization of the mix? .
|
|
#
?
Oct 18, 2011 06:08
|
|
|
incoherent posted:The Something Awful Forums > Discussion > Serious Hardware / Software Crap > Intel Roadmap Discussion: Ivy Bridge Roulette Spring 2012. K CPUs still have VT-x, the base virtualization that the last gen Cores had. They just don't have VT-d peripheral virtualization. E: VVV I'll take the old poo poo 
Factory Factory fucked around with this message at 06:49 on Oct 18, 2011 |
|
#
?
Oct 18, 2011 06:44
|
|
|
Intel Roadmap Discussion: Bulldozer =  Those TDP numbers for the performance chips, hot drat. I may get one just because there's every chance that the power draw will pay for itself over the lifetime of the system, and I can write off the expense.
|
|
#
?
Oct 18, 2011 06:47
|
|
|
VR-Zone also has an article saying that the upcoming Sandy Bridge-E (LGA2011) will only launch with two of its eight cores disabled. Apparently they aren't able to hit acceptable clockspeeds with all eight cores enabled within the allotted TDP. The second stepping of the chip should enable them to launch an uncut product. I wouldn't mind one of those 150W TDP processors, imagine how effective Turbo must be?
|
|
#
?
Oct 18, 2011 07:06
|
|
|
|
| # ? May 8, 2024 21:04 |
|
|
I'm going to have a really hard time not upgrading my desktop (E8400) and server (Q9450) with some new Ivy Bridge goodness... I just want the power savings really.
|
|
#
?
Oct 18, 2011 10:25
|
|
