Penryn Benchmarks (alledgedly) posted in China

[the_vorlon]

These from a China based site

No warrantty expressed or implied

3.33 ghz quad on 1333 fsb versus 2.93 ghz Quad on 1066 fsb

http://techgage.com/viewimg/?img=/articles/intel/beijing_idf_07/idf_beijing_01.jpg&desc=Intel Penryn

http://techgage.com/articles/intel/beijing_idf_07/idf_beijing_01.jpg

Divx + 211% => 85% "clock for clock"
Half life + 41% => 24% "clock for clock"
Cinebench + 25% => 10% "clock for clock"
Photoshop + 15% => 1% "clock for clock"

DivX obviously uses one of the new SSE "N+1" instructions, while on Halflife, 24% seems to indicate that the new 16 bit Radix divider works as advertized.

Not "stunning" increases, but solid gains, and considering that 3.33 is likely NOT the highest speed bin, certainly an optimistic looking release considering we are still 8 months out from release date.

 

[1Tanker]

8O 8O 8O holee friggen carp 8O ( shaking my head at AMD) My gawd dewd thats insane. I may just ditch AMD after all. If these numbers are correct man I may take the cramer outlook; why even think about amd at all.?

Vern, are you insinuating that there's something fishy about those numbers?? :lol: :lol:

 

[1Tanker]

with a great figure like that i could consider proposing,jeezus tank, Whats the point of AMD if they dont have the goods?

These numbers are insane and put penny at 60% over AMD k8 ,or more.

Won't matter, if they can't release it before Q2 08 (BK)

 

[carlhungis]

ok, but i smell Q1. At a premium to avoid undercutting c2d.

I doubt they will wait until Q1. Intel is more concerned about stealing AMD's thunder than they are about undercutting C2D. I would imagine it will release at a price to still have the C2D's as an attractive option.

 

[carlhungis]

ok, but i smell Q1. At a premium to avoid undercutting c2d.

I doubt they will wait until Q1. Intel is more concerned about stealing AMD's thunder than they are about undercutting C2D. I would imagine it will release at a price to still have the C2D's as an attractive option.

I wanted to say that as well, but I held back.

Which is funny, because I didn't really mean it as an insult or anything. I think that Intel sees it as more beneficial for the company to release it at a time that would hurt AMD's launch. Not as some sort of a "kill AMD at all costs" type of manuver. My understanding is that Penryn will be able to be launched whenever Intel needs to launch it. If AMD sticks to their roadmap and releases in Q's 2, 3 and 4, then I would expect to see Penryn right there in Q3. If AMD slips a little, then I would expect Penryn to slide right along with it.

I am sure that AMD fans will see this as a shady tactic, but honestly... why wouldn't Intel do it? What possible reason would Intel have to hold off on releasing Penryn and let AMD gain market share back? If you think that AMD wouldn't do the same thing if given the opportunity, then you are kidding yourself.

 

[carlhungis]

If AMD doesnt play a damn good hand here quickly,all that will be left is the core AMD users. I just havent decided if i am core,especially with High k penny knocking on the door.

How much do we placate the poor planning and not just go for the best and dump the ideal; for me there is a breaking point.

come to the dark side............

 

[qurious69ss]

If AMD doesnt play a damn good hand here quickly,all that will be left is the core AMD users. I just havent decided if i am core,especially with High k penny knocking on the door.

How much do we placate the poor planning and not just go for the best and dump the ideal; for me there is a breaking point.

come to the dark side............

Unlike 2002... we have people watching.....que the old twighligt zone muzic............

 

[Egon]

ok, but i smell Q1. At a premium to avoid undercutting c2d.

I don't think they will wait that long. According to the Q1 CC they have already done a large portion of their capital expenditure for 45nm and they will want to start selling asap to pay back that cost and take advantage of the smaller process and the inherent cost savings it provides.

Egon

 

[BaldEagle]

If AMD doesnt play a damn good hand here quickly,all that will be left is the core AMD users. I just havent decided if i am core,especially with High k penny knocking on the door.

How much do we placate the poor planning and not just go for the best and dump the ideal; for me there is a breaking point.

Welcome to the Brood as Baron refers to anyone that would dare to criticize his precious AMD.

 

[RichPLS]

8O :?: :lol: 8)

 

[warezme]

similar and more detailed benchs are up at anandtech... but, I thought the whole thing was dissappointing.

I mean factor in the higher FSB, higher CPU Mhz and bigger cache and a some extra SSE4 extensions beta run on a couple of media programs and that does not leave a big improvement overall really.

If they had been like 40 - 45% across the board..., I would say awesome, a monster, like everyone else is yelling..., so far is....ehhh...

 

[edwuave]

wtf...how high can this high-k overclock.

 

[RichPLS]

It would be nice if it brought 4+GHz as the min easily achieved OC!

 

[darious00777]

My thought right now would be, how much energy is it using at the tested rating? If it's 65 watts or less, then I'd say that's pretty good news. If it's 40 watts with the quad core, then I'd say that would mean they're doing a lot better then the benches would show.

Then again Wolfdales will probably be 55-60 watt chips and Penryns 85-100 watts. Just an uneducated guess.

 

[warezme]

My thought right now would be, how much energy is it using at the tested rating? If it's 65 watts or less, then I'd say that's pretty good news. If it's 40 watts with the quad core, then I'd say that would mean they're doing a lot better then the benches would show.

Then again Wolfdales will probably be 55-60 watt chips and Penryns 85-100 watts. Just an uneducated guess.

That is a valid point. The last I read was that the quad core would be equal in wattage as the current dual core. Very much like AMD's claims of Barcelona. I'm not sure how that would equate with Penryn duals.

It will be a good chip no doubt and I will still buy a quad Penryn upgrade but I just don't get the ferver I'm reading in comparisons with the just released benchmarks.

 

[Major_Spittle]

8O 8O 8O holee friggen carp 8O ( shaking my head at AMD) My gawd dewd thats insane. I may just ditch AMD after all. If these numbers are correct man I may take the cramer outlook; why even think about amd at all.?

holee friggen carp is right Vern. And it will probably be released in numbers vs a paper launch and right around the same time as the K10 desktop processors.

Here is the scary secret, it is really just a shrink of Conroe other than the new instructions. Sure the HKMG should allow it to get above 4 gig. but this is merely going to be the slap in the face to AMD's K10.

Nahelam will come soon after Penryn. This is no dumb shrink and will be what makes AMD's share holders soil themselves and AMD's management to update their resumes and cash out their stock options. I mean Nahelam will be a real drawer cleaner for AMD.

I don't know if you remember way back to OCT/November when I was posting the doom and gloom about AMD. But this was what I was getting at. And obviously AMD's execution of 65nm is horrible, sounds like K10 might not scale well at all.

I fear AMD is going the way of Cyrix unless IBM pulls something out of its Azz to save AMD. Maybe TSMC will by them, or a front company for the Chineese government. These truely are dark days for Baron's ilk. Too bad for AMD, I liked a lot of their products in the day.

 

[darious00777]

It will be a good chip no doubt and I will still buy a quad Penryn upgrade but I just don't get the ferver I'm reading in comparisons with the just released benchmarks.

I can understand that. I keep wondering how much it would cost to get something significantly more powerful then my current set-up, and with the costs in the thousands, I don't see the need for it yet.

Still, the idea of a quad core running with the same power usage of a Conroe is fairly exciting to me.

 

[dasickninja]

Vern, I've got one thing to say...

You know you wanna...

 

[ltcommander_data]

Here, I think this will be informative:

http://www.bit-tech.net/hardware/2007/04/18/idf_spring_2007_benchmarking_penryn/2.html

I have to admit, Intel's level of disclosure, while it may not be perfect is still quite amazing and admirable. They are actually admitting where the breakdown in the 105% performance increase of Penryn over Kentsfield comes from. Essentially:

75% SSE4
15% Frequency
10% Architecture
5% FSB

Sadly when you look at that breakdown, it doesn't quite look that amazing, but Intel is actually admitting it. Still, while SSE4 support is a wild card, the other 30% will be exposable in most applications. You can probably expect a 5-10% average gain clock for clock from the new architecture and the FSB room, which is very decent for a refresh. Frequency is Intel's back-up plan and they will implement depending on how Barcelona turns out and how quickly AMD can ramp and increase clock speeds.

 

[reconviperone1]

I remember aamd took there time releasing the qfx benches and we all know how that story went. How in the hell do you compete with that? Amd better be very equal, cause i dont think they outdo that monster. Of course my braand spanking new x2-5600 could kick the pneryn, barcelona, and athena ass.LOL

 

[dasickninja]

what kind of cookies? 8O

The ones that make women have orgasms and make babies say their first words.

 

[RichPLS]

In like Flynn!

 

[dasickninja]

But the orgasms are longer and harder and the words are bigger.

 

[RichPLS]

[/jaw drops] 8)

 

[dasickninja]

I baked them myself. ^_^ [/proud]