Sign in with
Sign up | Sign in
Your question

Is it just me???

Last response: in CPUs
Share
April 18, 2007 6:07:29 PM

I somehow expected more.

I mean like 40 - 45% more across the board.

If you consider the higher FSB of Wolfdale compared to Conroe, the bigger cache and the faster CPU Mhz..., that does not leave much of extra in your face performance really.

I found the whole thing .... dissappointing.

More about : question

April 18, 2007 6:12:03 PM

Even a 1066MHz FSB doesn't seem to limit performance for desktop tasks and even the 4mb Conroes don't show a big improvement compared to 2mb Conroes. I don't think it's really disappointing at all, I'm impressed that they're demoing a high clocked 45nm chip so far from release.
April 18, 2007 6:50:04 PM

I think the increase is pretty good as it is. As cpu's get faster, it gets increasingly difficult to get large percentage increases.
Related resources
April 18, 2007 6:52:05 PM

It's a very good increase due to the fact there were no major architectural changes. Basically the Intel engineers shrunk Conroe, went "Hey guys, I think we can toss this other stuff in without a problem," and did it.
April 18, 2007 6:54:35 PM

Quote:
I somehow expected more.

I mean like 40 - 45% more across the board.

If you consider the higher FSB of Wolfdale compared to Conroe, the bigger cache and the faster CPU Mhz..., that does not leave much of extra in your face performance really.

I found the whole thing .... dissappointing.


It would have been disappointing if it was slower than the Conroe.
April 18, 2007 7:29:21 PM

Quote:
I somehow expected more.

I mean like 40 - 45% more across the board.

If you consider the higher FSB of Wolfdale compared to Conroe, the bigger cache and the faster CPU Mhz..., that does not leave much of extra in your face performance really.

I found the whole thing .... dissappointing.

939 => AM2

THAT'S disappointing
April 18, 2007 9:17:23 PM

Quote:
I somehow expected more.

I mean like 40 - 45% more across the board.

If you consider the higher FSB of Wolfdale compared to Conroe, the bigger cache and the faster CPU Mhz..., that does not leave much of extra in your face performance really.

I found the whole thing .... dissappointing.

939 => AM2

THAT'S disappointing

HA!, that is/was dissappointing. I suppose when the OC threshold begins to appear we'll all know how good it really is.
April 18, 2007 9:21:43 PM

But being that its still real early there can be some changes and performance increases made before they release it.
April 18, 2007 9:32:32 PM

Don't expect 3.33Ghz to be the highest clock speed either. The fact that it is clocking so well so far from release, on top of throwing in some IPC bones, is why people are excited. That, and the fact that we have yet to see anything like this kind of disclosure from the K10 camp. One more powerpoint slide from AMD trumpeting spec_fp2000 performance and I would say a defenestration would be very much in order.
April 18, 2007 9:46:01 PM

Quote:
One more powerpoint slide from AMD trumpeting spec_fp2000 performance and I would say a defenestration would be very much in order.


forgive me, I had to look that up..., defenestration...., but my boot up their you no what would have been my choice.
a b à CPUs
April 18, 2007 9:47:21 PM

The hardware performance of the Conroes/Penryn/Wolfdale are fine. It is the software that needs to take advantage of this. I mean seriously, how many apps are multi threaded?? Not many if you think about it. and adding SSE4 would also help in the long run.
a c 99 à CPUs
April 19, 2007 12:06:23 AM

Quote:
The hardware performance of the Conroes/Penryn/Wolfdale are fine. It is the software that needs to take advantage of this. I mean seriously, how many apps are multi threaded?? Not many if you think about it. and adding SSE4 would also help in the long run.


We are only now starting to see SSE2 become widespread, let alone SSE3. SSE4 will give the Penryn/Wolfdale chips a big boost in things that support it, but I have a hunch that Intel's demo included about every program that has SSE4 at the moment. The non-SSE4 apps show a 15-20% increase over the QX6800, which is the 12% clock speed plus a few percent for the larger caches.

It's going to be a nice chip, no doubt, but it's a slightly tweaked die shrink, no more and no less. That's what the tests show for the vast majority of apps. But that's what Intel said it was supposed to be and that's exactly what it is. It's good to see Intel hitting decent clocks on 45 nm this far ahead of ship date, and what I'd really like to see is the Schmoo plot for the chip. That would be rather telling as to the minimum capabilities of the process, so I highly doubt that Intel would leak that. AMD sure as heck wouldn't :D 
April 19, 2007 12:23:36 AM

I think it is just you... These are just engineering samples...

Genereally shipping version will perform better and overclock like a bat out of 4377!

April 19, 2007 1:22:52 AM

40 - 50% across the board from a die shrink with minor architectural tweaks? You must be dreaming. ;) 

This isn't Netburst to Core2 again, we probably won't see such a jump for a looooonggggg time.
April 19, 2007 1:28:51 AM

Under 3DMax, Intels top chips are well over 100% Faster than AMD's F6000~!!! :p  :lol: 
April 19, 2007 2:05:55 AM

Actually, they are almost exactly 100% faster. They should be as there are 4 cores vs 2 in a program that is actually multi threaded, and very scalable as well apparently. Very impressive to be sure.
a b à CPUs
April 19, 2007 2:08:10 AM

/Law of diminishing returns
April 19, 2007 12:30:30 PM

Quote:
40 - 50% across the board from a die shrink with minor architectural tweaks? You must be dreaming. ;) 

This isn't Netburst to Core2 again, we probably won't see such a jump for a looooonggggg time.


To hear everyone exclaiming so, I figured thats what the buzz was about. Turns out it was more moderate increase in performance. Its all good, just not on the level of what the hype seems to be.

And as far as the 40%, I don't think any one of us will have to wait that loooong since Nehalem(sp) is on the heals of Penryn and its more of a major overhaul to the Core architecture as opposed to a refresh of Conroe that Penryn appears to be.

Its still all good. I just think lots of people and the media loose perspective and start spouting descriptors like massive, monster, etc., to describe what is essentially a very good refresh.

No one has seen any complete benchs let alone OC potential limits yet? Folks may be in for a rudeness if they discover the 1333Mhz bus and low multiplier tied to the Penryn limits their current mobos to low OC'ing or none at all.
April 19, 2007 12:44:23 PM

That does leave plenty of room for Intel to issue faster chips at higher multipliers as the process improves. They get more $$$ as well. Not a bad plan.
April 19, 2007 12:58:34 PM

Quote:
That does leave plenty of room for Intel to issue faster chips at higher multipliers as the process improves. They get more $$$ as well. Not a bad plan.


I think you're on to something there. Intel thinks they sold too many e6400's that are running at 3.2 gigs now, so they give you a chip that is less overclockable via the FSB. Now if you want a faster chip you have to go to the next price point for the better mutiplier. :wink:
!