Official Desktop Penryn Discussion Thread

Printable View

Show 100 post(s) from this thread on one page

07-18-2007, 02:01 PM
Movieman

Quote:

Originally Posted by red

July 15 was supposed to be Penryn NDA break according to TGDaily but that turned out to be a rehash of the June 24th FSB1333 articles plus the QX6850.. And that June 24th date was in turn supposed to be on July 22nd along with the price cuts, and now July 22nd is Penryn leak day. Huh? BTW, I thunk I saw that 2.33GHz Yorky elsewhere. Could be the same person, I personally don't think so though..

I talked with this guy and he seems very straight and honest to me.
I don't know if his cpu's were on that other post you read.
Let's wait and see what gets posted on sunday the 22nd..;)
07-19-2007, 02:39 AM
WilliamE

I did not have the NDA date, and thought it was OK to post. After a PM from another user, giving me a lecture on the NDA, I had to remove it.
I did post a few pictures on in the threat I asked moderator to remove. (One stock with cpu-z, one with the CPU laying in my hand, and one of my testbench)
I am sorry I got some of you exited, but I promise I‘ll make a new one when the NDA is off.

And BTW. I have never shown these pictures/screenshots anywhere else but on this forum :)
07-19-2007, 02:46 AM
informal

Quote:

Originally Posted by WilliamE

I did not have the NDA date, and thought it was OK to post. After a PM from another user, giving me a lecture on the NDA, I had to remove it.
I did post a few pictures on in the threat I asked moderator to remove. (One stock with cpu-z, one with the CPU laying in my hand, and one of my testbench)
I am sorry I got some of you exited, but I promise I‘ll make a new one when the NDA is off.

And BTW. I have never shown these pictures/screenshots anywhere else but on this forum :)

Thanks for sharing those,but stay low until the 22nd :).
07-19-2007, 03:46 AM
Jacky

Quote:

Originally Posted by WilliamE

I did not have the NDA date, and thought it was OK to post. After a PM from another user, giving me a lecture on the NDA, I had to remove it.
I did post a few pictures on in the threat I asked moderator to remove. (One stock with cpu-z, one with the CPU laying in my hand, and one of my testbench)
I am sorry I got some of you exited, but I promise I‘ll make a new one when the NDA is off.

And BTW. I have never shown these pictures/screenshots anywhere else but on this forum :)

damn can't wait until the 22nd (if that's the day the nda is lifted, now please don't tell me it isn't :( )

Yorkfield xtreme edition gets priced, 999$ (digitimes)
yeah "motherboard makers confirm" we can expect 45nm extreme editions for christmas :)
07-19-2007, 04:19 AM
Torchon

I'm looking at a roadmap on XbitLabs.

http://www.xbitlabs.com/images/cpu/c...50/roadmap.jpg

Are mainstream and performance Yorkfields really coming out in Q1'08? I was really looking
forward to some cheap Yorkfield action and had it all planned to buy a whole new PC at the
end of the year. Big dissapointment. :/

And BTW, the next Intel lineup is unfathomable. Yorkfields are desktop 45nm processors, right?
What are Penryns and Wolfdales? Could someone clear this up for me please?
07-19-2007, 04:24 AM
Shintai

We will get Yorkfield and server versions in Q4.

Anyway:
Yorkfield=Quadcore Desktop
Wolfdale=Dualcore Dekstop
Penryn=Dualcore Mobile
07-19-2007, 04:38 AM
nemrod

Quote:

Originally Posted by WilliamE

I did not have the NDA date, and thought it was OK to post. After a PM from another user, giving me a lecture on the NDA, I had to remove it.
I did post a few pictures on in the threat I asked moderator to remove. (One stock with cpu-z, one with the CPU laying in my hand, and one of my testbench)
I am sorry I got some of you exited, but I promise I‘ll make a new one when the NDA is off.

And BTW. I have never shown these pictures/screenshots anywhere else but on this forum :)

Thank you ;)

That lets you a lot of time to run a lot of benchmarks :D
07-19-2007, 04:51 AM
WilliamE

I have never mentioned the 22nd as the date they are no longer under NDA.
Personally I don’t think it’s the 22nd, I think it’s later…
07-19-2007, 04:54 AM
Movieman

Quote:

Originally Posted by WilliamE

I have never mentioned the 22nd as the date they are no longer under NDA.
Personally I don’t think it’s the 22nd, I think it’s later…

Yes, there is some confusion on that and I think it's due to Intel having it's price cut on that day..Some may have mixed the 2 up.
07-19-2007, 04:54 AM
Jacky

what about this new thread about a yorkfield ES
http://www.xtremesystems.org/forums/...d.php?t=151920

can we trust the numbers/the source?
do we have to expect "only" 5-10% ipc improvement? :down:
oh and the chips seems to run a little hot and I hoped for even lower vcore ... 5x °C at idle if the readings are correct (stock cooling?)

EDIT: HUH? so you don't even know when the NDA ends? movieman said something about the 22nd so I thought it may be 22nd but I asked you, because you never mentioned the date... aaaah so no benchmarks coming?! just give us some hints or anything then...
07-19-2007, 04:58 AM
Shintai

It will "only" be 5-10% faster in IPC. Remember, its mainly just a dumb dieshrink with SSE4+50% more cache+Radix+Supershuffle.

For the heat part, I dont think you can rely on those numbers. Also Intel demoed 3.33ghz and run around with 3.2Ghz Quads at crysis and some warcraft3 stuff etc with those. And they use a stock cooler. BIOS thermal sensor numbers for the CPU is a lottery event. Voltage dont seem to be so wrong tho. But could change. 1.2V I presume vs current 1.35V
07-19-2007, 05:01 AM
coffeetime

Which would be the better buy, Yorkfield Extreme or equivalent server version?

Like the current Q6600 is the exact same thing as the Xeon 3220 Quad, just a different name and supposedly binned higher because it's a server grade CPU?

I'm so hurting for this new PC lol, still on 5yr old 2.4c Northwood :rofl::shakes:
07-19-2007, 05:03 AM
Shintai

The 3000 series.
07-19-2007, 05:32 AM
Jacky

Quote:

Originally Posted by Shintai

It will "only" be 5-10% faster in IPC. Remember, its mainly just a dumb dieshrink with SSE4+50% more cache+Radix+Supershuffle.

For the heat part, I dont think you can rely on those numbers. Also Intel demoed 3.33ghz and run around with 3.2Ghz Quads at crysis and some warcraft3 stuff etc with those. And they use a stock cooler. BIOS thermal sensor numbers for the CPU is a lottery event. Voltage dont seem to be so wrong tho. But could change. 1.2V I presume vs current 1.35V

QFT

however I saw 10% increase in cinebench and 20% in HL2 over at anandtech in their preview and I was optimistic and hoped to see some more 10%++
eventually it will depend on the individual workload, for instance SSE4 will show "some more" improvements than 10%...

now contributing some news:
arstechnica summing up some of the news about penryn
07-19-2007, 07:15 AM
Shadowmage

Quote:

Originally Posted by Jacky

QFT

however I saw 10% increase in cinebench and 20% in HL2 over at anandtech in their preview and I was optimistic and hoped to see some more 10%++
eventually it will depend on the individual workload, for instance SSE4 will show "some more" improvements than 10%...

now contributing some news:
arstechnica summing up some of the news about penryn

The anandtech comparison was comparing a 2.93ghz Conroe vs 3.33ghz Wolfdale I believe, so it wasn't clock for clock.

5-10% IPC improvement with some significant outliers is more than reasonable.
07-19-2007, 01:04 PM
Donnie27

Benchmark Yorkfield Performance Advantage
3DMark '06 V1.1.0 Pro CPU.... (score) ..: 21.8%
3DMark '06 V1.1.0 Pro Overall.. (score) ..: 7.6%
Mainconcept H.264 Encoder (seconds) .: 18.0%
Cinebench R9.5 (CPU test).................: 24.9%
Cinebench R10 Beta (CPU test)...........: 25.5%
HL2 Lost Coast Build 2707 (fps) ...........: 37.3%
DivX 6.6 Alpha w/ VirtualDub 1.7.1 (seconds) 111%

But the performance difference in more than the 10% speed difference. Lord help the competition if optimzations are used:shocked: Sorry guys, IMHO, 3DMarks anything is a waste of benchmarking time.
07-19-2007, 02:06 PM
informal

Don't pay much attention to those numbers Donnie,since it is in controlled env. and under unknown test conditions(which optimizations were used,ie SSE etc).

But ,you can look forward to some 5-10% just thanks to bigger L2 and other already mention arch. improvements.SSE is another point,and will need apps that support it to take advantage of it.
07-19-2007, 02:13 PM
Shintai

Quote:

Originally Posted by informal

Don't pay much attention to those numbers Donnie,since it is in controlled env. and under unknown test conditions(which optimizations were used,ie SSE etc).

But ,you can look forward to some 5-10% just thanks to bigger L2 and other already mention arch. improvements.SSE is another point,and will need apps that support it to take advantage of it.

All the SW and versions was shown, no fancy stuff with custom optimization besides DiVX (But you can download that publicly).

Intel have proven more than once that the benchmarks holds water. Unless you wanna argue against the early Conroe benches too.

SSE enabled games seems to gonna like the supershuffe engine.
07-19-2007, 03:27 PM
Jacky

Quote:

Originally Posted by Shintai

All the SW and versions was shown, no fancy stuff with custom optimization besides DiVX (But you can download that publicly).

Intel have proven more than once that the benchmarks holds water. Unless you wanna argue against the early Conroe benches too.

SSE enabled games seems to gonna like the supershuffe engine.

yeah but they probably choose the best benchmarks anyway (those showing 10%+ ipc wise :D ). btw does superpi use SSE?
07-19-2007, 05:29 PM
Donnie27

Quote:

Originally Posted by informal

Don't pay much attention to those numbers Donnie,since it is in controlled env. and under unknown test conditions(which optimizations were used,ie SSE etc).

But ,you can look forward to some 5-10% just thanks to bigger L2 and other already mention arch. improvements.SSE is another point,and will need apps that support it to take advantage of it.

I don't see any reason for them to fudge any of those results and even they gave a disclaimer about optimized software.

If you take away or try to factor the in 10% difference (just for scale, an example only) you're talking 8 to 27% clock for clock, not counting the optimized app. I'm sure we all use optimized free apps, right?

I agree with Shintai, after the Conroe pretty much proved itself, Intel would be stupid to blow any of the credit they've earned.:up:
07-19-2007, 05:30 PM
Donnie27

Quote:

Originally Posted by Shintai

All the SW and versions was shown, no fancy stuff with custom optimization besides DiVX (But you can download that publicly).

Intel have proven more than once that the benchmarks holds water. Unless you wanna argue against the early Conroe benches too.

SSE enabled games seems to gonna like the supershuffe engine.

QFT!
07-19-2007, 05:38 PM
Donnie27

Quote:

Originally Posted by Jacky

yeah but they probably choose the best benchmarks anyway (those showing 10%+ ipc wise :D ). btw does superpi use SSE?

But if something AMD launched was 10% faster, most of you guys would say AMD spanked Intel.:D
07-20-2007, 10:33 AM
red

http://www.hardspell.com/doc/hardware/39109.html
Translator says Yorkfield will be 2% of Q407, Q108 5%, Q208 6%. Wolfdale will be 13% Q108, 23% Q208. In summary, for desktop, 2% 45nm in Q407, 18% in Q208, 29% in Q308.
07-23-2007, 07:47 PM
Donnie27

Quote:

Originally Posted by red

http://www.hardspell.com/doc/hardware/39109.html
Translator says Yorkfield will be 2% of Q407, Q108 5%, Q208 6%. Wolfdale will be 13% Q108, 23% Q208. In summary, for desktop, 2% 45nm in Q407, 18% in Q208, 29% in Q308.

Try this?

http://www.hardspell.com/english/
07-27-2007, 06:27 AM
Jacky

Quote:

Originally Posted by red

http://www.hardspell.com/doc/hardware/39109.html
Translator says Yorkfield will be 2% of Q407, Q108 5%, Q208 6%. Wolfdale will be 13% Q108, 23% Q208. In summary, for desktop, 2% 45nm in Q407, 18% in Q208, 29% in Q308.

"With Intel set to launch 45nm quad-core (Harpertown) and dual-core (Wolfdale-DP) server processors in fourth quarter 2007, 45nm processors are scheduled rapidly increase in shipment proportions to account for almost 50% of total 2-way server shipments by second quarter 2008, according to sources at server makers."

digitimes

Show 100 post(s) from this thread on one page

All times are GMT -8. The time now is 01:29 PM.

XtremeSystems