Six Core AMD reviewed

[Praz]

i7 930 beat it at lower clock speed in 6 out of 8 tests.... Epic fail in my book. Amd needs new architecture to make an impact.

Nobody is really interested in reading your book. You post the same thing in every thread regarding the Thuban. You really need to find something else to post about.

[ccokeman]

i7 930 beat it at lower clock speed in 6 out of 8 tests.... Epic fail in my book. Amd needs new architecture to make an impact.

 

 

Ok lets see 8 threads vs 6 hhhmmmm Whats going to win in multi threaded apps? For what it is its a huge step up for AMD at a ridiculously low price point and provides a real upgrade path with improved performance over the 4 core AMD chips.

[dr_bowtie]

Agreed...

 

and you CANNOT compare clock speed from one part to the next...they are two totally different companies with two totally different architectures...

 

anyone who would compare and Intel versus an AMD at teh same clock speeds is missing the whole point

[joel.monteiro]

im still confused.

[El_Capitan]

Agreed...

 

and you CANNOT compare clock speed from one part to the next...they are two totally different companies with two totally different architectures...

 

anyone who would compare and Intel versus an AMD at teh same clock speeds is missing the whole point

Wait... what's the point I'm missing comparing them at the same clock speeds? The whole point in comparing them at the same clock speeds is to monitor how much of a difference their architecture makes.

 

1. We all know 4 cores at 4.2 GHz is going to be faster than 4 cores at 3.9 GHz.

2. If we know that, then we want to know how an i7 920 at 3.9 GHz without Hyperthreading compares to a Phenom II x4 965 at 3.9 GHz.

3. If it's practically the same when running benchmarks, then that's good to know. We'd then want to know what a difference an i7 920 at 3.9 GHz WITH Hyperthreading compares.

4. We then move onto testing an i7 930 at 4.2 GHz with and without Hyperthreading (4 cores, then 4 cores with 8 threads) against a Phenom II x6 1090T (6 cores) at 4.2 GHz.

5. The question is... does Hyperthreading make an impact? From my experience, for video encoding, it does. What about database multi-threaded operations? Which one performs better?

 

First, there's going to be quite a few reviews comparing CPU's at stock speeds. The ones that win benchmarks are the ones with higher clock speeds with more cores. Duh! We don't need MORE of those tests because we already know the outcome. The sole reason to test CPU's at the maximum stable 24/7 scenario for the lowest overclockable CPU is to gain insight on the price per performance ratio. At places like "OVERCLOCKERSclub", people overclock. Why give reviews at stock speeds? Plus, most reviews give GAMING benchmarks! Why? There's hardly a bump in performance when having over 4 cores from 3 cores, there's not going to be a difference from 4 to 6, especially when most games still aren't multi-threaded. Another thing I hate... most of these reviews don't state whether the i7's have Hyperthreading enabled.

 

My guess, from all the benchmarks that matter that I've been reading... i7 920 d0 wins.

http://www.anandtech.com/show/3674/amds-si...055t-reviewed/6 - i7 975 at 3.33 GHz beats the 1090T at 3.2 GHz. Since we know the i7 920 can be overclocked to 3.33 GHz with minimal effort, those 2 extra cores aren't adding anything useful.

 

http://www.anandtech.com/show/3674/amds-si...055t-reviewed/7 - again, same results.

 

http://www.techspot.com/review/269-amd-phe...055T/page7.html - different review, same results.

 

http://www.legitreviews.com/article/1289/7/ - different review, same results.

http://www.legitreviews.com/article/1289/9/ - same results.

 

http://www.tomshardware.com/reviews/amd-ph...0fx,2613-7.html - close to same results.

 

http://www.bit-tech.net/hardware/cpus/2010...black-edition/5 - it barely beats the i7 930 at stock while the 1090T is overclocked

http://www.bit-tech.net/hardware/cpus/2010...black-edition/6 - same.

 

I'll stop there. I have 2 Intel systems (dual XEON server and desktop), and 3 AMD systems (desktops). I was really hoping the 1090T would be worth getting, but I'd rather upgrade to an i7 930 and buy another x58 motherboard (if I needed another power system)... but I'm still hoping AMD steps it up a notch. The 1090T is overrated in my book.

 

EDIT: On a side note, the Phenom II x3 720 BE's as budget gaming workstations are great! At Microcenter, a good AM3 motherboard and 720BE combo costs from $105 - $155, depending on the motherboard you get. Can't beat that price!

Edited April 30, 2010 by El_Capitan

[joel.monteiro]

Nobody is really interested in reading your book. You post the same thing in every thread regarding the Thuban. You really need to find something else to post about.

in (games) at a high res like 1680x1050 there is a very big difference.

[El_Capitan]

I think what the reviewers and moderators of this site are really saying is... even though the 1090T can't beat the i7 920, it'll be less expensive.

 

Let's see: i7 920 on sale HERE for $205 (at Microcenter, the i7 930 for $205, or i7 920 for $170)...

1090T on sale at Newegg for $305...

 

Newegg:

x58 board for $160 - $200...

AM2+/AM3 board for $60 - $100...

 

Seems pretty even to me.

[sdy284]

in (games) at a high res like 1680x1050 there is a very big difference.

lol @ 1680x1050 being high res

[The Smith]

First, I am not going to re-explain why stock clocks comparisons are fine, as it was already explained above.

 

My guess, from all the benchmarks that matter that I've been reading... i7 920 d0 wins.

http://www.anandtech.com/show/3674/amds-si...055t-reviewed/6 - i7 975 at 3.33 GHz beats the 1090T at 3.2 GHz. Since we know the i7 920 can be overclocked to 3.33 GHz with minimal effort, those 2 extra cores aren't adding anything useful.

 

http://www.anandtech.com/show/3674/amds-si...055t-reviewed/7 - again, same results.

 

http://www.techspot.com/review/269-amd-phe...055T/page7.html - different review, same results.

 

http://www.legitreviews.com/article/1289/7/ - different review, same results.

http://www.legitreviews.com/article/1289/9/ - same results.

 

http://www.tomshardware.com/reviews/amd-ph...0fx,2613-7.html - close to same results.

 

http://www.bit-tech.net/hardware/cpus/2010...black-edition/5 - it barely beats the i7 930 at stock while the 1090T is overclocked

http://www.bit-tech.net/hardware/cpus/2010...black-edition/6 - same.

You fail to prove your point.

 

Let's see; From the 20 graphs you've linked to, Here is where the Phenom II loses against the 920 / 930 (depending which one they used) :

-3dsMax

-iTunes

-Cinebench R10 SINGLE thread (It does beat it here though but I am not going to count it as you didn't link to these graphs. And in multi-thread, it wins on both charts.)

-GIMP

-Multi-tasking test from Bit-Tech

-overall score from Bit-Tech

 

Then you also link to a graph where it loses against the 930 in Handbrake, but oddly enough it does beat it in your other link here. So which one tells a lie? I'd say the Phenom II has the win, but once again I am not going to count it as you didn't link to Neoseeker.

 

Another contradiction in your post is TMPGEnc. Here you show it losing against anything, but here it beats everything but Gulftown, and is well ahead of Bloomfield.

 

Considering contradictions a tie, this means the Phenom II X6 loses only 6 times on 20 against Bloomfield. In my book, this is what I call a win.

Edited April 30, 2010 by The Smith

[El_Capitan]

First, I am not going to re-explain why stock clocks comparisons are fine, as it was already explained above.

 

 

You fail to prove your point.

 

Let's see; From the 20 graphs you've linked to, Here is where the Phenom II loses against the 920 / 930 (depending which one they used) :

-3dsMax

-iTunes

-Cinebench R10 SINGLE thread (It does beat it here though but I am not going to count it as you didn't link to these graphs. And in multi-thread, it wins on both charts.)

-GIMP

-Multi-tasking test from Bit-Tech

-overall score from Bit-Tech

 

Then you also link to a graph where it loses against the 930 in Handbrake, but oddly enough it does beat it in your other link here. So which one tells a lie? I'd say the Phenom II has the win, but once again I am not going to count it as you didn't link to Neoseeker.

 

Another contradiction in your post is TMPGEnc. Here you show it losing against anything, but here it beats everything but Gulftown, and is well ahead of Bloomfield.

 

Considering contradictions a tie, this means the Phenom II X6 loses only 6 times on 20 against Bloomfield. In my book, this is what I call a win.

 

First, I am not going to re-explain why stock clocks comparisons are fine, as it was already explained above.

Stock comparison are not fine. Please read my whole reply. Those that buy enthusiast CPU's to use it for heavy multi-tasking applications and don't overclock it to their full potential is just wasting money and time they don't know they're wasting. The whole point in me being here on this site is because it's an OVERCLOCKER's CLUB. Not a STOCKCLOCKER's CLUB. If I'm wrong about this site, then I should just move on over to bit-tech.net, Overclock.net or Overclockers.com. I already moved here from Tom's Hardware. I just want to pick one place and stick with it.

 

You fail to prove your point.

You fail to see my point, but I don't need to put it in big fonts. I'm trying to be logical and civil. I don't want to argue with a child, it's just not worth the effort.

 

1. Please re-read my statement: i7 975 at 3.33 GHz beats the 1090T at 3.2 GHz. Since we know the i7 920 can be overclocked to 3.33 GHz with minimal effort, those 2 extra cores aren't adding anything useful.

2.

Then you also link to a graph where it loses against the 930 in Handbrake, but oddly enough it does beat it in your other link here. So which one tells a lie? I'd say the Phenom II has the win, but once again I am not going to count it as you didn't link to Neoseeker.

- Refer to point #1 to re-read my statement. The i7 930 loses at stock speeds, but the 1090T loses to the i7 975 at 3.33 GHz.

3.

Another contradiction in your post is TMPGEnc. Here you show it losing against anything, but here it beats everything but Gulftown, and is well ahead of Bloomfield.

- Refer to point #1 to re-read my statement... again. The i7 930 loses at stock speeds, but the 1090T loses to the i7 975 at 3.33 GHz.

 

Considering contradictions a tie, this means the Phenom II X6 loses only 6 times on 20 against Bloomfield. In my book, this is what I call a win.

Please... understand what I'm talking about first before you make a fool out of yourself. If you're only going to read and quote part of my reply, that doesn't help.

 

If the 1090T really does beat the i7 920 - each to it's full potential in a stable 24/7 system, it doesn't need to be marketed in big fonts. Logical assumptions and benchmarks should be enough. Trust me, I WANT to want the 1090T. I'm just not sold on it. Prove me otherwise, and I'll convert. If at all anything, I'd buy the damn thing to just run my own benchmarks to prove you wrong.

 

One last thing. Look here again:

http://www.bit-tech.net/hardware/cpus/2010...black-edition/5

http://www.bit-tech.net/hardware/cpus/2010...black-edition/6

 

See how the i7 930 at 4.3 GHz just pwns the 1090T at 3.87 GHz? Now, I imagine the scores of the 1090T would be better if it could match the clock speed of the i7 930, but they'd need to update that review if they could get the 1090T to a higher overclock. I have to give bit-tech.net credit for the best 1090T review.

Edited April 30, 2010 by El_Capitan

[joel.monteiro]

1680x1050 is not a very low res dude u can call it medium.

[The Smith]

Stock comparison are not fine. Please read my whole reply. Those that buy enthusiast CPU's to use it for heavy multi-tasking applications and don't overclock it to their full potential is just wasting money and time they don't know they're wasting. The whole point in me being here on this site is because it's an OVERCLOCKER's CLUB. Not a STOCK CPU USER's CLUB. If I'm wrong about this site, then I should just move on over to bit-tech.net, Overclock.net or Overclockers.com. I already moved here from Tom's Hardware. I just want to pick one place and stick with it.

 

1. Please re-read my statement: i7 975 at 3.33 GHz beats the 1090T at 3.2 GHz. Since we know the i7 920 can be overclocked to 3.33 GHz with minimal effort, those 2 extra cores aren't adding anything useful.

2. - Refer to point #1 to re-read my statement. The i7 930 loses at stock speeds, but the 1090T loses to the i7 975 at 3.33 GHz.

3. - Refer to point #1 to re-read my statement... again. The i7 930 loses at stock speeds, but the 1090T loses to the i7 975 at 3.33 GHz.

Ok I'll try again.

 

Overclockers (like us members here) are only a small percentage of the readers. Most people do not overclock their CPUs. These are sold at a speed that is guaranteed to work. They want to know the performance they are getting out of the box. Of course if you tune the Bloomfield, it's going to beat it! But what tells you it is going to be easy? Nothing. You're making an assumption here. Everyone knows that every chip is different. And who knows? It might require a great increase in voltage, thus reducing its lifespan. You might even fry it!

 

For manufacturers, increasing the frequency is just another way to increase performance, as adding more cores or adding more threads is. In Intel's case, yes, their architecture is more powerful for a given clock. So what? It's like saying that a Porshe will go faster than a firetruck with a given engine. You need to equip the firetruck with a more powerful engine if you want it to reach the same speed.

Edited April 30, 2010 by The Smith