Woooooahhh: new OFFICIAL PII speed record

[kid41212003]

They do overclock better than the first generation of Phenoms, the average 24/7 stable overclock on air with a Phenom II 920 / 940 will be at approx 3.75GHz, with some good chips reaching or even surpassing the 4GHz mark by a bit, and some "duds" stopping at 3.6GHz or so.

I expect Phenom II would match the Yorkfield performance clock vs clock or slower ~5%.
I expect the Phenom II BE would reach 3.7GHz under air , and 3.5GHz for non-BE CPU, running 24/7.

My prediction of few days ago was quite close...

 

[BenchZowner]

The Phenom 2 did 33 MB in two seconds more than the i7 at only 30Mb. How exactly is that a fair benchmark?

2.3 Mbps i7
2.2 Mbps Phenom



100K difference per second standardized. Not much, a simple miscache by the CPU will have greater effects, and well within what a normal variance is.

Allow me to educate you
WinRAR displayes the "final" average compression rate ( resulting compression rate ), and the current compression rate is being updated every second after the "final" ( resulting ) rate is shown.
The run time before the resulting compression rate is displayed is lower than what you see in the screenshots, it took me some time to get the process paused.

From WinRAR's documentation ( in case you don't feel like trusting me ):

This command generates random data, which contain specially introduced redundancy increasing load to processor and memory. Then data are passed through RAR compression and decompression algorithms and output of decompression algorithm is compared with source data. If any difference found, WinRAR reports "Errors found - Yes" in the command window. Also WinRAR displays a size of processed data and compression speed, current and resulting, in kilobytes per second. You may use the resulting speed value to compare RAR performance in different conditions. For example, "Benchmark" command may be helpful, when you need to choose a new computer and wish to know which one will compress data faster. Only the general compression algorithm in "Best" mode with 4096K dictionary is called, all additional filters and algorithms are disabled, so it measures performance of core RAR compression on worst case like data.

Resulting speed is displayed when at least 10MB of source data are processed, collected statistics is enough to get an accurate result and current speed changes are near zero. After the resulting speed has been set, it is not changed more. Though source data are random, their redundancy level and other parameters are always the same. So this command will report practically the same current speed regardless of execution time, provided that system load is not changing.

 

[Steevo]

Practically and "is" are two different things.


Thats practically your child she is carrying, you had sex wither her, and she had sex with a few others.

VS

That IS your child she is carrying!


See, big difference. How about a 200MB zip file extraction?

 

[KBD]

It's not a bad processor, and is a worthy upgrade for people with AM2 boards and older processors.
Unfortunately it ain't the big bang that we were expecting and hoping to see.
But they don't really need a new architecture to catch up or even surpass Intel, they just need to enhance or redesign some portions of the CPU.

yea, i never sed it was bad, i sed it was a sweet CPU and in fact i'm considering getting it myself when the new DDR3 boards come out as i much rather get an AMD then Intel. But i cant say i agree on the architecture, unless you can explain this. I mean K10b (or whatever they call the Deneb) is just a revised K10 which in turn is a revised K8. The latter has been around since 2003 and while AMD were on top they made few enhancements to it and their CPU development stagnated as they became overconfident. So basically no new architecture came to light since 2003. Now they are saying that they will have a new architecture based on 32nm in 2011. But its too late, intel will have the first 32nm CPUs by Q4 of 09 or Q1 of 10. The botton line is that AMD had nothing new since 2003 in terms of introducing a new architecture. Now what else can they do to the current architecture to be competitive? It seems like they already done all the tweaks they could and Deneb is the final result of these enhancements.

 

[BenchZowner]

Practically and "is" are two different things.


Thats practically your child she is carrying, you had sex wither her, and she had sex with a few others.

VS

That IS your child she is carrying!


See, big difference. How about a 200MB zip file extraction?

You're clearly forgetting that:

1) The HyperThreading Technology is disabled, as a result you're seeing a small difference between the CPUs, while it can be way bigger.

2) Once again what you can't seem to understand is that:
Once I leave the Core i7 reach the 15s mark, it's already at 35 and sometimes at 36.

The extraction process is less dependent on the CPU, but the Core i7 is still faster.

 

[PaulieG]

I think it is clear to most of us that i7 is just better in most, if not all applications than PII. Not speaking for anyone else, what I want to know is how it competes across the board with penryn, and how much better it is than PI in terms of overclocking and overall performance. I'd also love a screenie of BZ's 940 at 4.0ghz.

 

[cdawall]

the thing about that i7 being faster in this test is still the HT actually working in one thing the same will happen with video editing etc but tests that don't utilize HT will be much more effective in showing the "true" differences between these tests

 

[BenchZowner]

The rumors regarding the CPUs that some overclockers had and the others AMD used in their demonstrations were cherry-picked samples is looking more and more possible now.

We have a total of 3 Phenom II's here, two 940s and one 920.
On air our best 940 did 3.89GHz "fully stable" ( 6 hours prime95 ), our second 940 did 3.77GHz "fully stable" ( 5 hours prime95 ) and the 920 stopped at 3.71GHz "fully stable" ( 6 hours prime95 ) and benchable at 3.95GHz, would rarely pass SuperPi 1M at 4G on air.

We're done testing the second 940 on LN2, nothing to write home about, 5.5GHz benchable.
We're going to take another break now and grab some more pizza to eat
Then we're going to give the other 940 a try with LN2.

Another person that I know and trust has also tested another retail 940 ( he works with a hardware distributor ) on LN2 and maxxed out at 5.4GHz ( stable in 3D Marks, PC Marks, CineBench 10, SuperPi 32M, wPrime, etc ) and ran some light benches ( SuperPi 1M, Hexus PiFast ) at 5.52GHz.

the thing about that i7 being faster in this test is still the HT actually working in one thing the same will happen with video editing etc but tests that don't utilize HT will be much more effective in showing the "true" differences between these tests

If I for example run Video Editing and Encoding programs, and time is valuable to me, I'd pick a Core i7, it's simply faster, with and without HT.
Of course if the tasks that I'll be running benefit from HT I will enable it, any reason I should not ?

In reality, if our budget allows us to buy the best CPU we grab it, if our budget is limited we take a look at our options, compare them in tasks that we usually do and pick the most suitable one.
We're about to receive some Octo-Core Bloomfields ( Core i7, 8 cores, 16 threads ) E.S. what is AMD going to do about it ? Hm... don't know to be honest.
Life ain't fair, and if you're going to compare something you use all of its features if they're beneficial to you

 

[PaulieG]

I want some screenies!

 

[HTC]

New OC WR for Phenom II

http://valid.canardpc.com/show_oc.php?id=466047

Question: what's with those odd timings? Shouldn't it be like 5-5-5-15/18 or 6-6-6-18/21 or something?

More records here:

http://valid.canardpc.com/records.php

 

[BenchZowner]

I'll give you guys more screenies later on today or tomorrow.
We have two PC's on LN2 right now, and we can't waste time & nitrogen to find the screenies and upload them.

Gotta get back in action asap, rotation time

 

[Solaris17]

I'll give you guys more screenies later on today or tomorrow.
We have two PC's on LN2 right now, and we can't waste time & nitrogen to find the screenies and upload them.

Gotta get back in action asap, rotation time

haha well thanks for the time and effors so far....and i didnt see anyone welcome you to tpu...so welcome

 

[PaulieG]

I'll give you guys more screenies later on today or tomorrow.
We have two PC's on LN2 right now, and we can't waste time & nitrogen to find the screenies and upload them.

Gotta get back in action asap, rotation time

Hey, what was your vcore on the 3.89ghz on air run?

 

[HTC]

haha well thanks for the time and effors so far....and i didnt see anyone welcome you to tpu...so welcome

Dude: he's been registered since October of 2007

Stands to reason that he was welcomed then, no?


@ BenchZowner: many thanks for all your posts in this topic!

 

[Raiderman]

Dude: he's been registered since October of 2007

Stands to reason that he was welcomed then, no?


@ BenchZowner: many thanks for all your posts in this topic!

He's been a member since 07, but his first post was actually in this thread.


So a welcome is not out of order.

 

[Solaris17]

Dude: he's been registered since October of 2007

Stands to reason that he was welcomed then, no?


@ BenchZowner: many thanks for all your posts in this topic!

good point i was going by post count and what i saw in this thread......intresting i dont remember you bench and iv been here quite awhile

 

[BenchZowner]

haha well thanks for the time and effors so far....and i didnt see anyone welcome you to tpu...so welcome

Thanks for the welcome

Well it's not really a effort, we're just having fun here

Just loaded a image to an old Maxtor 80GB IDE drive and run wPrime 1024m on a Q6600 ( Kentsfield ) at 3.1GHz to compare it with our run with the Phenom II 920.
I used a DDR2 board to make things more equal ( if that can be said, since Kentsfield is 2 generations old now for Intel, and the Phenom II is a "brand new" chip that's supposed to face the Core i7 or the Yorkfield at least, since it cannot take on Core i7 no matter what ).

The Kentsfield completed the 1024m run in 408.528s and the Phenom II in 390.539s.

 

[Solaris17]

Thanks for the welcome

Well it's not really a effort, we're just having fun here

Just loaded a image to an old Maxtor 80GB IDE drive and run wPrime 1024m on a Q6600 ( Kentsfield ) at 3.1GHz to compare it with our run with the Phenom II 920.
I used a DDR2 board to make things more equal ( if that can be said, since Kentsfield is 2 generations old now for Intel, and the Phenom II is a "brand new" chip that's supposed to face the Core i7 or the Yorkfield at least, since it cannot take on Core i7 no matter what ).

The Kentsfield completed the 1024m run in 408.528s and the Phenom II in 390.539s.

that isnt much of a diffirence.....could you do me a favor seeing as alot of Q6600's will do between 3.4-3.6 would you mind clocking that high and seeing if it can overtake the phenom?

 

[HTC]

He's been a member since 07, but his first post was actually in this thread.


So a welcome is not out of order.

I stand corrected

Have you dudes noticed the new record? 6120 MHz

 

[ascstinger]

Hey, what was your vcore on the 3.89ghz on air run?

I would be interested to know this as well. I have a loop that'll be cooling my PII, but I dont want to run a boatload of voltage just to hit the magic 4.0.

 

[cdawall]

http://valid.canardpc.com/show_oc.php?id=466047

Question: what's with those odd timings? Shouldn't it be like 5-5-5-15/18 or 6-6-6-18/21 or something?

More records here:

http://valid.canardpc.com/records.php

its an old tweak method i used it on my ancient s754 rig trying to push the bus speed that extra notch. you loose performance but it lessens the stress on the IMC

 

[HTC]

I would be interested to know this as well. I have a loop that'll be cooling my PII, but I dont want to run a boatload of voltage just to hit the magic 4.0.

While this question has merit, another much more important one begs to be asked: what temps on idle and full load with that speed and Vcore?

 

[Raiderman]

The Kentsfield completed the 1024m run in 408.528s and the Phenom II in 390.539s.

Just completed a run on my 9850@2809 mhz
1024M-432.6 secs.

Big improvement in the phenom II

 

[Raiderman]

While this question has merit, another much more important one begs to be asked: what temps on idle and full load with that speed and Vcore?

Yes I am interested in the temps also

 

[cdawall]

since they used the old loose subtimings trick got me thinking are the IMC's on these chips weaker?