New Data for Piledriver

[Obsoleet]

Besides, even among us enthusiasts, how many of us already say faster CPU isn't needed anymore?

Almost all of us.
I do more than most do with their machine (encoding/decoding, fileserver, gaming, streaming hub) and I've had more than I need since the release of my Q9450. I'd like an upgrade, but dont need one. Waiting for Haswell and to see if AMD can get Bulldozer CPU speeds with 5870 integrated graphics performance. Be a great all-in-one machine that suits my needs, and surpasses 95% of the users out there as well.

 

[Shaydza]

Almost all of us.
I do more than most do with their machine (encoding/decoding, fileserver, gaming, streaming hub) and I've had more than I need since the release of my Q9450. I'd like an upgrade, but dont need one. Waiting for Haswell and to see if AMD can get Bulldozer CPU speeds with 5870 integrated graphics performance. Be a great all-in-one machine that suits my needs, and surpasses 95% of the users out there as well.

It's gonna be a while for a 5870 class graphics to show up on a CPU. Maybe we can get close to a 5770 soon though. Perhaps on Vishera

 

[Olikan]

In desktops, Llano isn't clocking so low, so the gain for Trinity is much less.

trinity desktop is supposed to have 3.8-4.2 GHz

 

[T_Yamamoto]

trinity desktop is supposed to have 3.8-4.2 GHz

 

[jacktesterson]

I honestly think AMD is doing its best to close the gap with what they have to work with. Too much invested in BD.

I'm quite impressed if that slide is true.

 

[pelov]

That seems to be uniform across segments though, whether Intel or AMD, graphics or CPU. The gains on high end looks to be "ehh".

I'm not sure that's the case with the on-die graphics. We've seen some awesome gains with SB > IB and Llano was amazing and Trinity will be a 50% gain over that. I still feel the RCM within Piledriver may be benefiting the mobile parts moreso than the desktop, and though you're right that we see this with Intel as well, it may be a bit more profound in the desktop Piledriver parts. The 25% gain looks to be true when comparing equal-TDP laptop parts but the desktop... i dunno. It'll still be a healthy increase but I wouldn't expect the 25% gains we see in the 35W mobile chips.

 

[Olikan]

I'm not sure that's the case with the on-die graphics. We've seen some awesome gains with SB > IB and Llano was amazing and Trinity will be a 50% gain over that. I still feel the RCM within Piledriver may be benefiting the mobile parts moreso than the desktop, and though you're right that we see this with Intel as well, it may be a bit more profound in the desktop Piledriver parts. The 25% gain looks to be true when comparing equal-TDP laptop parts but the desktop... i dunno. It'll still be a healthy increase but I wouldn't expect the 25% gains we see in the 35W mobile chips.

it might be diferent at overclocking, but the top end trinity will run at 3.8-4.2Ghz at 100W TDP...with a IGP of 66W TDP @ 40nm...

 

[DeeDot78]

The next Gen bulldozer speculation is a 4.2stock - 4.5 turbo Vishera, due to mesh technology. These are the kind of clocks they were hoping for with first gen bulldozers. If you think AMD isnt trying to get back in Enthiusiats CPU market you are mistaken. That press release was about buying time and taking off pressure. They are trying to right the ship. When they move to next socket after AM3+, they'll be more competitive.

Note: These are mobile numbers.

 

[joshhedge]

The next Gen bulldozer speculation is a 4.2stock - 4.5 turbo Vishera, due to mesh technology. These are the kind of clocks they were hoping for with first gen bulldozers. If you think AMD isnt trying to get back in Enthiusiats CPU market you are mistaken. That press release was about buying time and taking off pressure. They are trying to right the ship. When they move to next socket after AM3+, they'll be more competitive.

Note: These are mobile numbers.

I get close to those scores on my almost 4 year old Core 2 Duo MBP...

 

[DeeDot78]

I get close to those scores on my almost 4 year old Core 2 Duo MBP...

With trinity level integrated graphics, in a laptop, with a small tdp? I dont think so. No baiting, but thats not a comparision to make.

 

[joshhedge]

With trinity level integrated graphics, in a laptop, with a small tdp? I dont think so.

Stuck with my 8600M GT 512mb, damn. I don't care about tbp though. I presume trinity offers greater graphics performance, haven't compared with my GPU yet.

 

[pelov]

I'm assuming multi-threaded? If Trinity is performing that well at 2.3ghz (plus turbo? which I'm assuming might get quite high) that would be quite amazing.

it might be diferent at overclocking, but the top end trinity will run at 3.8-4.2Ghz at 100W TDP...with a IGP of 66W TDP @ 40nm...

Yea, the desktop parts should clock very high

I'm more and more interested in seeing just how these chips pan out. It might turn out that AMD wasn't nucking futs and there actually was some hope in the BD design and clock speed goals, albeit it took an extra 2 years to get there.

 

[Vesku]

That A10-5700 listing looks promising at 65W.

 

[Shaydza]

That A10-5700 listing looks promising at 65W.

Agreed. We may have the next ultimate HTPC chip coming our way with this one.

 

[IntelUser2000]

I'm not sure that's the case with the on-die graphics.

Exactly, on-die graphics isn't high end.

Geekbench also isn't the best benchmark. I can find a Core i3 that scores better than that Core i5: http://browse.geekbench.ca/geekbench2/search?utf8=✓&q=2310m

It scores low as 3900 and high as 5900.

 

[pelov]

Exactly, on-die graphics isn't high end.

Oh yea, there are some large strides to be made there still and I'd expect some great gains over the next couple generations but that too will stagnate just as everything else has :|

 

[ElFenix]

That's not "Total fail", the two words you're looking for are unbelievably impressive. I can't think of the last time I saw that kind of gain without a die shrink.

yorkfield -> nehalem?

clarkdale -> sandy?

 

[Arzachel]

Oh yea, there are some large strides to be made there still and I'd expect some great gains over the next couple generations but that too will stagnate just as everything else has :|

CPU performance has stagnated not because we've hit a technological barrier, but because more performance does amazingly little for the most part. It would be very hard for me to see the difference between a SB clocked at 3ghz and one at 4ghz when simply browsing the web, and a SSD just does much more for the typical user experience. People want smaller/lighter/sleeker because you can actually feel the change as opposed to a clock speed bump. iGPU performance has been the focus for both AMD and Intel because most tasks requiring a GPU have a baseline performance, under which it becomes really unpleasant, if not impossible. You can feel the difference between over or under 30fps in games far more vividly than when a video is taking 20mins longer to render.

Hardware development isn't stagnating, companies are just moving their resources to ares which make the largest impact.

 

[DeeDot78]

For anyone interested, phoronix did a article on the new gcc 4.7 compiler. Here are numbers for intel 3960x and AMD FX 8150. AMD is the top chart with bdver1 as bulldozer optimized code. Proves how important optimization of software/compilers is.

Bulldozer gets a nice boost from new compilers. Here the link to article.

http://www.phoronix.com/scan.php?page=article&item=amd_fx8150_compilers&num=1

 

[richaron]

^ that's nice