• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

New Data for Piledriver

D

DeeDot78

Member
6295945_8174fda20c_m.png


More like a glimpse. Should be improved even further for Trinity desktop Version.
 
They still have not caught up with Sandy Bridge i5-2500k. Which -in less than a month- will be replaced by its IVB equivalent with about 10% CPU improvement.

But AMD have advantage on the iGPU side, so losing to IVB by, say, 15% on CPU side while outperforming IVB iGPU by 40% might be an acceptable tradeoff for some OEMs and users.

Would be interesting to see how this plays out and if AMD can stall their marketshare erosion. On the notebook side I think situation should be even more fascinating, would be hilarious if Intel popularized the 'ultrabook' format and AMD took the lion's share.
 
meloz said:
They still have not caught up with Sandy Bridge i5-2500k. Which -in less than a month- will be replaced by its IVB equivalent with about 10% CPU improvement.

But AMD have advantage on the iGPU side, so losing to IVB by, say, 15% on CPU side while outperforming IVB iGPU by 40% might be an acceptable tradeoff for some OEMs and users.

Would be interesting to see how this plays out and if AMD can stall their marketshare erosion. On the notebook side I think situation should be even more fascinating, would be hilarious if Intel popularized the 'ultrabook' format and AMD took the lion's share.
Click to expand...
+1
 
So ivb chips will have integer performance that is 50% greater, AND they will probably be clocked 10-20% higher for a total of 60-75% better integer performance. Total fail...
 
sm625 said:
So ivb chips will have integer performance that is 50% greater, AND they will probably be clocked 10-20% higher for a total of 60-75% better integer performance. Total fail...
Click to expand...

w0t?

The IVB desktop chips are clocked only 100mhz higher than their SB counterparts. The performance increase from SB > Ivy is essentially that single digit IPC bump and 100mhz. Where you're getting 60-75% I have no idea but you're wayyyyy off. If Trinity is pulling integer performance up by ~20% over the 4100 which has L3 and slightly lower clocks (Trinity lacks L3 cache) we'll probably be looking at a pretty damn good Trinity/Vishera Piledriver core. If the Vishera parts are the same clocks as the Trinity part there's a good chance we're looking at 15-25% increase from BD>PD *without a die shrink*. 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.
 
If they get 970 performance out of it then that's good enough. We've had enough cpu power since the i7 920 came out almost 3 years ago IMO. Personally I'd be more interested in a low power all in one solution from Amd that can run games at medium for all AAA games at 1080p.
 
For a desktop APU part without L3? Yea, it's awesome. Don't expect IPC levels to be the same as the Intel chips, but it seems they've gotten quite close to the Llano/Husky levels which was the intent with Bulldozer. If Vishera clocks mid/high 4ghz with Turbo then we'll be looking at quite a decent chip.
 
If this info is true and holds up in benchmarks, then i just might grab a Trinity processor for my secondary pc. 😉
 
OILFIELDTRASH said:
If they get 970 performance out of it then that's good enough. We've had enough cpu power since the i7 920 came out almost 3 years ago IMO. Personally I'd be more interested in a low power all in one solution from Amd that can run games at medium for all AAA games at 1080p.
Click to expand...

Agreed, my core 2 quad 9650 will have a very vey long productive life. I'd really perfer to see less power, less heat and I agree some kind of intergrated graphics that actually allows you to play games.
 
Hell I'd bet a lowly i3 2100 would have a damn long life for gaming provided there is a video card swap at some point.
 
Last edited:
If thats the performance increase, they still should have die shrunk thuban. They seem to be going in a dead end direction IMO.
 
pelov said:
w0t?

The IVB desktop chips are clocked only 100mhz higher than their SB counterparts. The performance increase from SB > Ivy is essentially that single digit IPC bump and 100mhz. Where you're getting 60-75% I have no idea but you're wayyyyy off. If Trinity is pulling integer performance up by ~20% over the 4100 which has L3 and slightly lower clocks (Trinity lacks L3 cache) we'll probably be looking at a pretty damn good Trinity/Vishera Piledriver core. If the Vishera parts are the same clocks as the Trinity part there's a good chance we're looking at 15-25% increase from BD>PD *without a die shrink*. 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.
Click to expand...

"Unbelievably impressive" to have gotten the IPC back to where the Phenom 2 was several years ago???
 
frozentundra123456 said:
"Unbelievably impressive" to have gotten the IPC back to where the Phenom 2 was several years ago???
Click to expand...

If you ignore this part...

If the Vishera parts are the same clocks as the Trinity part there's a good chance we're looking at 15-25% increase from BD>PD *without a die shrink*. 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.
Click to expand...

The good news is that there's clearly something salvageable in the Bulldozer design. I'm not happy with the CMT approach either because moar coars doesn't equate to more performance but this chip is looking to be what Bulldozer should have been in the first place: Nehalem performance. The even better news here is that the Nehalem performance is coming from an APU that's got VLIW4 graphics attached to it limiting TDP and likely clock speeds. If Vishera clocks even higher, coupled with that L3 cache it should be pretty damn good.
 
pelov said:
If you ignore this part...



The good news is that there's clearly something salvageable in the Bulldozer design. I'm not happy with the CMT approach either because moar coars doesn't equate to more performance but this chip is looking to be what Bulldozer should have been in the first place: Nehalem performance. The even better news here is that the Nehalem performance is coming from an APU that's got VLIW4 graphics attached to it limiting TDP and likely clock speeds. If Vishera clocks even higher, coupled with that L3 cache it should be pretty damn good.
Click to expand...

Watch it pelov, you are starting to sound hopeful.

You know what hope brings?

Disappointment.

😛

(FWIW, I agree with you.)
 
As someone who hated nearly everything about Bulldozer even I have to give credit where credit is due 😛 If they are in fact squeezing that much out of the Trinity APU then they've hit a home run. It won't catch Intel's SB nor IB in CPU performance but it should still be competitive and absolutely whoop it in graphics/gaming.

I hope it is true. Don't think anyone wants to see AMD slip even further behind. They still have 2 or 3 of these to go, though
 
This is a good step forward. I wonder if this is implementing the resonant clock mesh they talked about before?

Anyways, if they can get the IGP about 60% faster than Llano like projected, and the power characteristics are good, they'll have one heck of a mobile chip. Priced correctly, they'll really erode Intel's market share there.
 
rpjkw11 said:
Is it possible to "disable" the onboard video and add a discrete graphics card to the A series?

Bob
Click to expand...

Why would you want to?? The onboard video is the only thing outstanding about Llano. You can also add a low end discrete card to work with the onboard, but I dont know if they have the kinks worked out yet. Initially the asymetric crossfire did not work well at all.
 
These numbers are a move in the right direction. Include the new mesh tech, IPC increase, and high(er) clocks, and we should have a nice Bulldozer replacement. I think this is only the beginning of the tweaking of the design. And I wish people would stop thinking they should just shrink thuban, NOT going to happen. They spent a boatload on this design in R&D(whether u like the dosing or not), they have to have a return of some-type.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top