Intel Accelerates Launch of Ivy Bridge by Almost 1 Week

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.
[citation][nom]jsc[/nom]That is the same prerelease hype that BullDozer came with.[/citation]

Yeah, and fanboys were screaming bloody murder if someone suggested that Bulldozer wasn't going to live up to the hype.
 
hmmm I'm seeing lots of chatter around the web about Ivy Bridge being a failure, because it missed it's 77W TDP target and is now rated at 95W TDP...the same as Sandy-bridge, which indicates a big manufacturing problem since Ivy Bridge is a die shrink.

If it wasn't for PCI Express 3.0 x16 and native USB 3.0 support I'd avoid Ivy Bridge completely, unfortunately I've already purchased a Z77 Motherboard and a GTX 680...seem silly to limit those with a Sandy-bridge because of a"little heat." lol
 
Since modern games offload more of the video processing to the GPU than ever before, from a gaming standpoint, it seems that given the same GPU(s) and a given resolution, any of the new processors and underlying architectures since 2008 (1366/X58) have only given us around 0-5 more fps. I wonder if we could see another architecture comparison 1366 through LGA 2011 and a GTX 680 or two in SLI. We still aren't seeing bottlenecks with the X58 architecture.

Though Intel continues to pursue efficiency, it seems they've continued to repackage similar gaming performance for the last few years. I know the processor isn't all about gaming, but I'm not hell-bent on updating any of my PCs to shave milliseconds off compiling code.

The things that really drive me to update my rig are those giving me a substantial increase in my gaming performance. I don't see IB as one of those things.
 
I finally got fed up with my old Phenom II setup last weekend and got a 2500k, even though my buddy urged me to wait, thinking that Ivy Bridge was going to be about the same price for higher performance. Now I'm glad I didn't wait. I'm happy, running way faster at the same temps, and it'll be even better once I upgrade the stock cooler. Before reading this article, I was thinking that in a couple months maybe, I'd give my current rig to one of my sons and put together a new build based on Ivy Bridge, but with talks of "substantially higher" prices, the platform just won't be attractive to me for quite a while.
 
Let us hope that Intel will see that it is in their interest to not allow computer OEM's to write custom Intel HD graphics drivers for Intel HD 4000 integrated graphics! I have a low end Toshiba laptop with a Intel first generation core i3 processor, and I have tried to update the INTEL HD 2000 graphics drivers, only to be told, at the Intel HD graphics website, that my graphics drivers can not be updated by Intel because Toshiba, the OEM for my laptop, has written customized graphics drivers! Toshiba has never, and probably never will, update the drivers, so I will not be able to use my Wings3d software with the Toshiba laptop, as I can not highlight any of the selected Faces of the 3D models I have created using Wings3D, and I need to update the graphics drivers to fix this problem! I bought the Toshiba in hope of having an inexpensive laptop to use the free open source graphics software, such as Wings3D, but I have had to get an ASUS laptop, and Wings3D works just fine, and I get regular Intel HD graphics drivers updates with my ASUS laptop. One question I will always ask when I am buying a laptop with Intel HD graphics GPUs is “Does this Laptop use Intel generic HD graphics drivers” , and If the answer is no, I will not buy the laptop! I will never again purchase a Toshiba laptop!
 
Luckily, my 3 year old computer will still run Diablo 3 perfectly, so I get to save aloooooooooooooot of money for the next two years at least, still.
 
[citation][nom]jkflipflop98[/nom]Remember back when technically inclined people posted on THG? People that actually knew what they're talking about? Those were the days.[/citation]
Awww come on, you can't make such a generalisation. At least point out what people are doing wrong... please? :)
 
Temperature measurements program are indicating wrong values, because they don't understand the architecture yet.

It's impossible to have the CPU run at 98 degrees and not throttling severely / burning itself.

Definitely a bug.
 
Toshiba has never, and probably never will, update the drivers, so I will not be able to use my Wings3d software with the Toshiba laptop, as I can not highlight any of the selected Faces of the 3D models I have created using Wings3D, and I need to update the graphics drivers to fix this problem! I bought the Toshiba in hope of having an inexpensive laptop to use the free open source graphics software, such as Wings3D, but I have had to get an ASUS laptop, and Wings3D works just fine, and I get regular Intel HD graphics drivers updates with my ASUS laptop.

Using a laptop with Intel graphics and Wings 3D? Seriously, son? Your complaints are fruitless - you've got not one but two wrong tools for the job. Head over to CGTalk.com or something, dig around the internet and do more research.

It's like someone complaining that their Phillips screwdriver won't work in flat-head screws.
 
[citation][nom]killerclick[/nom]Thanks AMD for falling so much behind and letting Intel rape customers with their prices.[/citation]
This is more of intel's fault than AMDs. You can bet that AMD wishes to be a much larger company. But go back to the mid 2000s. AMD's selling a faster, cheaper and cooler running CPU during the P4/Netburst era. Eventually, people start buying them over intel's.

But Intel has put road-blocks up preventing companies like Dell from selling AMD systems.

So, AMD doesn't sell as many CPUs, doesn't make the money which would have been used to pay for R&D. R&D suffers... causing AMD to lose market position. intel has some very very deep pockets.
 
The moral of the story is: do not expect great graphics processing from Intel if Intel is unwilling to maintain the graphics drivers for the GPU hardware that they build, and Allowing OEM's to own the update process for an Intel product just does not work!
 
wat about the notebook IB? will they finally deliver unprecedented performance all the while giving more battery life and no heat issues??
 
[citation][nom]CaedenV[/nom]If you asked me last summer I was planning on waiting for an IB CPU and 600 series nvidia card. Then along came a project that demanded that I upgrade my rig last November, and I could not be happier! The 600 series looks like it does not work so well for productivity loads, and the IB platform does not seem to be much better than SB, but with a higher price tag to pay for that massive die shrink and bug GPU that would be unused in my rig. All in all I am quite happy with my purchases, and likely will not be upgrading for 4 years, which should put me on a skylake/skymont CPU.Still, while not a massive improvement it is still progress to move to a new die size and better !/W. I mean look at AMD, with their new line they are actually worse on a !/W basis, and for gaming they are simply different, trading blows between the new and the old chips. IB may not be what we were expecting, but it is still progress none-the-less.[/citation]

Radeon 7750 is on par with the 6750 and 5750 in performance for much lower power usage (evidenced by it having no PCIe connectors due to a 55w TDP versus the 5750/6750 requiring a six pin connector for much greater power usage). In gaming, the 7970 uses significantly less power than the 6970 despite having the same TDP and it has far less idle power usage, all while being two tiers of performance ahead of the 6970.

If you meant AMD's CPUs and not AMD's graphics, well that is a different story. AMD's FX-4xxx CPUs match Phenom II in performance and performance per watt (more or less) due to their higher clock frequencies making up for their lower IPC. For example, at stock, AMD's fastest processor for gaming (excluding BF3) is the FX-4170. Overclocked, the Phenom II 960T is AMD's best gaming processor and unlocked plus overclocked, it's also AMD's best for gaming in BF3.

So, they didn't go up in performance besides at stock, but performance per watt isn't too different. It's bad for AMD, but it's not as bad as you suggested. Piledriver should have some huge improvements anyway, so AMD will probably make up a lot of lost ground. If they do it again with Steamroller, then Haswell won't be good enough to knock AMD out of mid-range gaming yet and really, that's what matters for AMD right now. The FX quad cores are good enough for a 6870 (FX-4100) and 6950 (FX-4170). Low/mid-end gaming are where the money really is and AMD is still good enough for that. It's really only at the top of the mid end and in the high end that AMD is no longer a viable option.

Piledriver might change that, especially if Trinity can do CF with it's more powerful than Llano IGP and a 7750/7770. If AMD can manage this much (they probably can), then they will do very well despite not being the performance winner in the high end.
 
[citation][nom]jsc[/nom]That is the same prerelease hype that BullDozer came with.[/citation]

Yes, but we now know Bulldozer's weaknesses. Here's a run down of how easily Piledriver could hit Nehalem and beyond performance if AMD does this right without even changing the CPU architecture much at all.

1. Bulldozer CPUs designing methods were poor. The CPU die designs were completely computer generated designs. Although this is faster to create and easier than a partially hand-optimized designs, it is usually about 20% larger, uses 20% more power, and about 20% slower than a design that has been hand-optimized. All other CPUs are entirely hand-designed (or much more common nowadays) or partially hand-designed (performance-critical parts are hand-designed and optimized). Fixing this would bring up performance significantly, drop the cost to manufacture significantly, and decrease power usage significantly.

2. Bulldozer has high latency cache. Improving the cache could also improve performance greatly because the cache is highly performance-critical.

3. Bulldozer (and really, all other AMD and most Intel CPUs excluding the Sandy/Ivy Bridge CPUs and maybe the Nehalem CPUs) has very inefficient memory controllers. They have about 25% less bandwidth than the Sandy Bridge controller does at the same memory frequency and same number of active memory channels.

Fixing those three problems alone, without improving the architectural problems, would offer a huge performance increase while decreasing power usage. It wouldn't be as efficient as Sandy Bridge yet, but it would meet or beat Nehalem. Fixing architectural problems such as increasing the issues of the front-end (Bulldozer shares four between the two cores in each module) and such could bring it up to Sandy Bridge's performance and beyond.

Will AMD manage this? Who knows. At least what needs to be done is obvious. With Bulldozer, we had little to go on besides speculation and rumor until it landed and we saw the reviews about how bad it is. I'm willing to give AMD the benefit of the doubt in that Piledriver will most certainly be a great departure from Bulldozer, but how far, I can not say. I can only say how far it can theoretically go, based on what I know of it's problems.
 
[citation][nom]ubercake[/nom]Since modern games offload more of the video processing to the GPU than ever before, from a gaming standpoint, it seems that given the same GPU(s) and a given resolution, any of the new processors and underlying architectures since 2008 (1366/X58) have only given us around 0-5 more fps. I wonder if we could see another architecture comparison 1366 through LGA 2011 and a GTX 680 or two in SLI. We still aren't seeing bottlenecks with the X58 architecture.Though Intel continues to pursue efficiency, it seems they've continued to repackage similar gaming performance for the last few years. I know the processor isn't all about gaming, but I'm not hell-bent on updating any of my PCs to shave milliseconds off compiling code. The things that really drive me to update my rig are those giving me a substantial increase in my gaming performance. I don't see IB as one of those things.[/citation]

An FX-4100 bottlenecks anything more than a Radeon 6870 in gaming. An FX-6100 compared to an i5-2500K@4.5GHz, both with dual 6950s, the i5 can have more than double the performance of the FX-6100 system despite having the same graphics system. This was proven in a Tom's SBM that compared the i5-2400 to the FX-6100 and just comparing the i5-2400 to the i5-2500K @4.5GHz in other benchmarks, you get the picture. Going by direct comparisons of the 6100 and 2500K, the same is true. In fact, it was so bad that an i5-2400 with a single 6950 beat the FX-6100 with dual 6950s most of the time. Having older/weaker CPUs is all okay with mid/low end setups, but getting into the upper mid-range and high end, not a chance.

X58 is beaten by LGA 1155 somewhat. That it is six fast cores against four faster cores makes them trade blows slightly and be very close. X79 is significantly better than X58, but only marginally better than LGA 1155.
 
[citation][nom]belardo[/nom]This is more of intel's fault than AMDs. You can bet that AMD wishes to be a much larger company. But go back to the mid 2000s. AMD's selling a faster, cheaper and cooler running CPU during the P4/Netburst era. Eventually, people start buying them over intel's.But Intel has put road-blocks up preventing companies like Dell from selling AMD systems.[/citation]

The size of the company means very little, since the actual manufacturing is usually handled by other companies.
ARM is tiny compared to Intel or AMD, yet they're the uncontested leader of the mobile CPU segment.
Intel, as big and rich as it is, wasn't able to make a GPU to compete with relatively small nVidia and ATI. Meanwhile, Hewlett Packard is a huge company, bigger than Intel, AMD, nVidia and ARM combined, yet their innovation is nonexistent.

If AMD doesn't have enough resources to produce a competitive CPU, they should either abandon that segment or abandon other segments to concentrate on the few products they can do well. There are no prizes for a distant second place. The reason AMD is this situation is that they dropped the ball on Bulldozer, and their management was too stupid and incompetent to pull the plug in time.
 
[citation][nom]killerclick[/nom]The size of the company means very little, since the actual manufacturing is usually handled by other companies.ARM is tiny compared to Intel or AMD, yet they're the uncontested leader of the mobile CPU segment.Intel, as big and rich as it is, wasn't able to make a GPU to compete with relatively small nVidia and ATI. Meanwhile, Hewlett Packard is a huge company, bigger than Intel, AMD, nVidia and ARM combined, yet their innovation is nonexistent.If AMD doesn't have enough resources to produce a competitive CPU, they should either abandon that segment or abandon other segments to concentrate on the few products they can do well. There are no prizes for a distant second place. The reason AMD is this situation is that they dropped the ball on Bulldozer, and their management was too stupid and incompetent to pull the plug in time.[/citation]

Actually, Intel was able to make a graphics card that could compete with Nvidia and AMD in the high end (Larrabee), but Intel decided that the market didn't have high enough profits and switched the graphics into a professional compute card instead of a gaming card. Look up Knights Corner or whatever the current version is called.

AMD can't abandon the CPU market because doing so would create an illegal monopoly for Intel and Intel would go as far as to pay AMD to stay in the market. Besides that, AMD's Piledriver is supposed to be a big leap over Bulldozer and even Phenom II in performance. Bulldozer is not a bad architecture, the CPU's were simply designed with poor methods and have poor cache coupled with poor memory controllers.

Bulldozer based FX CPUs are compeltely computer designed. All other CPUs are entirely or partially hand designed because a completely computer generated design always performs far below the others while using far more power (computers are not good for optimizing a design, just for rushing it). The cache has huge latencies and is slow. The memory controller is about 25% less efficient with memory frequencies than the Sandy Bridge controller (IE at 1333MHz dual channel, the FX will have about 25% less bandwidth than the Intel).

Fixing those basic problems will skyrocket Bulldozer's performance (and lower power usage greatly) without even fixing architectural problems. Going to the architectural problems, increasing the four issue-wide front end of each module to a six or eight issue wide front end should help tremendously. I could go on and on. Bulldozer really isn't bad, the CPUs that use it were simply designed poorly and their poor cache and poor memory controller exacerbate the problem.

AMD's processors are still great for low/mid end gaming and that's where the money is. Also, Intel is about three times as big as HP, so no, HP is not larger than Intel. Intel is far larger than HP, Nvidia, AMD, and ARM combined. ARM also isn't tiny. ARM is larger than Nvidia (who is even larger than AMD)*.

* all measurements based on market cap numbers from Yahoo Finance April/16/2012 at 10:01 PM Eastern USA time.
 
Status
Not open for further replies.