Hardware Secrets Forums


Go Back   Hardware Secrets Forums > Miscellaneous > Content Comments



Reply
 
Thread Tools Search this Thread Display Modes
Old 10-12-2011, 03:40 PM   #1
Hardware Secrets Team
Administrator
 
Join Date Nov 2004
Posts: 5,578
Hardware Secrets Team is on a distinguished road

Default AMD FX-8150 vs. Core i5-2500K and Core i7-2600K CPU Review

There has been a new article posted.

Title: AMD FX-8150 vs. Core i5-2500K and Core i7-2600K CPU Review
URL: http://www.hardwaresecrets.com/artic...PU-Review/1402

Here is a snippet:
"After so much postponing, AMD is finally releasing its new processor series based on the "Bulldozer" architecture, the FX, featuring models with four, six, or eight processing cores. Let's test the to..."

Comments on this article are welcome.

Best regards,
Hardware Secrets Team
http://www.hardwaresecrets.com
Hardware Secrets Team is offline   Reply With Quote
new Sponsored Links

Old 10-12-2011, 11:16 PM   #2
C.W.
Senior Member
 
C.W.'s Avatar
 
Join Date Jan 2011
Posts: 425
C.W. is on a distinguished road

Default

While the general consensus is that Bulldozer failed to meet expectations, oddly your review is one of the more positive reviews (relatively speaking) that I've seen. A couple other sites had the FX-8150 being outpaced by the older Phenom IIs 4x which is pretty awful.

One of the main differences I noticed with your gaming benches is that you prefer high res + low effects, while other sites tended to go low res + high effects or low res + low effects (or high + high, but that leads to obvious GPU bottlenecks). It seems like bulldozer does better with high res and reduced effect settings over low res and high effect settings.

Also, your iTunes was different in that you encoded into mp3 verses ACC (in which Bulldozer did much worse).


I hope AMD works to solve Bulldozer's weaknesses. While it's still overpowered by Intel, perhaps they can mature the architecture and redeem themselves.
C.W. is offline   Reply With Quote
Old 10-13-2011, 12:24 AM   #3
Korot
Junior Member
 
Join Date Feb 2010
The Netherlands
Posts: 9
Korot is on a distinguished road

Default

I agree with C.W., this review seems to be a bit more positive. It's still a shame though that Bulldozer didn't preform the way it was supposed to be doing.

BTW, did you also run benchmarks on the overclocked version, to indicate how much a difference it makes?

Lastly, on page 2 you stated that the FX8150 has 8MB L2 cache per module, however, it is in my understanding that it has 8MB Total, so 2 MB per module.

Regards,
Korot
Korot is offline   Reply With Quote
Old 10-13-2011, 07:15 AM   #4
Gabriel Torres
Administrator
 
Gabriel Torres's Avatar
 
Join Date Oct 2004
USA
Posts: 4,415
Gabriel Torres is on a distinguished road

Default

Thanks for the nice words. That is why it is so important to have several different publications performing different tests.

In fact, I thought our results were wrong, especially the Cinebench test, because it was unbeliveable that a four-core CPU could be faster than an eight-core CPU for 3D rendering.

And thanks for your research regarding games.

Quote:
Originally Posted by Korot View Post
Lastly, on page 2 you stated that the FX8150 has 8MB L2 cache per module, however, it is in my understanding that it has 8MB Total, so 2 MB per module.
You are right, I am fixing this right away.

Gabriel.
__________________
Editor-in-Chief
Hardware Secrets
http://www.hardwaresecrets.com
Gabriel Torres is offline   Reply With Quote
Old 10-13-2011, 12:19 PM   #5
Rafael Coelho
Super Moderator
 
Rafael Coelho's Avatar
 
Join Date Jun 2008
Posts: 177
Rafael Coelho is on a distinguished road

Default

Maybe current applications aren't optimized for eight-core CPUs.
__________________
There´s nothing as practical as a good theory.
Rafael Coelho is offline   Reply With Quote
Old 10-13-2011, 04:35 PM   #6
nikolarun
Junior Member
 
Join Date Oct 2011
Posts: 1
nikolarun is on a distinguished road

Default

Great positive review. In this day and age most software developers can only afford to optimise for one platform, and this is Intel. I truly believe there is more performance hidden in the chip. Maybe some benchmarks under Linux with amd/intel/third party compliers will show the chip's true potential.

Just a small point about the article, it will be helpful if all the CPU's kept the same colour in the graphs. It will make comparing them much easier.
nikolarun is offline   Reply With Quote
Old 10-13-2011, 10:02 PM   #7
C.W.
Senior Member
 
C.W.'s Avatar
 
Join Date Jan 2011
Posts: 425
C.W. is on a distinguished road

Default

You might try and perform a memory test where you compare a FX-8150 with two DDR3 sticks @ 1333, 1600, and 1866 (and higher?) verses say a Phenom II X6.

I noticed that your tests occasionally scored higher than other sites benches without any real reason unless it's the GPU or RAM (a lot sites used less than the native 1866). It would be interesting to see if Bulldozer benefits from faster RAM compared to its predecessor. Previously, most testing showed that faster RAM provided very little gains in terms of performance (with the exception of very memory dependent applications).
C.W. is offline   Reply With Quote
Old 10-14-2011, 09:33 AM   #8
Gabriel Torres
Administrator
 
Gabriel Torres's Avatar
 
Join Date Oct 2004
USA
Posts: 4,415
Gabriel Torres is on a distinguished road

Default

Thanks for the memory test suggestion. This is a common question. We've collected data for the A8 "APU" with DDR3-1333 vs. DDR3-1866 and we still need to post the results, but it is less than 3% improvement in overall performance. Meaning that most users won't get any benefit for upgrading the memory... Let's wait and see...
__________________
Editor-in-Chief
Hardware Secrets
http://www.hardwaresecrets.com
Gabriel Torres is offline   Reply With Quote
Old 10-15-2011, 04:59 PM   #9
Cheetos
Senior Member
 
Cheetos's Avatar
 
Join Date Jun 2008
Brazil
Posts: 426
Cheetos is on a distinguished road

Default

Quote:
Originally Posted by Rafael Coelho View Post
Maybe current applications aren't optimized for eight-core CPUs.
Cinebench is, some guys said that the other are too.

Quote:
Originally Posted by C.W. View Post
You might try and perform a memory test where you compare a FX-8150 with two DDR3 sticks @ 1333, 1600, and 1866 (and higher?) verses say a Phenom II X6.

I noticed that your tests occasionally scored higher than other sites benches without any real reason unless it's the GPU or RAM (a lot sites used less than the native 1866). It would be interesting to see if Bulldozer benefits from faster RAM compared to its predecessor. Previously, most testing showed that faster RAM provided very little gains in terms of performance (with the exception of very memory dependent applications).
http://www.madshrimps.be/articles/ar...#axzz0V1roTcmY

Quote:
Originally Posted by Gabriel Torres View Post
Thanks for the memory test suggestion. This is a common question. We've collected data for the A8 "APU" with DDR3-1333 vs. DDR3-1866 and we still need to post the results, but it is less than 3% improvement in overall performance. Meaning that most users won't get any benefit for upgrading the memory... Let's wait and see...
According to other reviews, the faster memory only improves the performance of Llano when the "graphic part" is used.

The Bulldozer isn't bad by itself, if AMD were able to get higher clock speeds (~4,5 GHz) they're get an earning FX codename processor, but they don't, and the power consumption is high, much higher than Sandy Bridges.

Coupled with that, the prices are too high, AMD failed by creating a Hype before the launch, making everyone thinks that the Bulldozers would be outstanding, and this chips are 6-9 months delayed, all problems they already made with Barcelona/Agena some years ago.

Now, they're trying to sell a "xx gigahertz chip, with XX MiBs of cache, XX cores, XXXX MHz memory support" instead of a fast, well priced and low power draw chip.

In fact, if they called it a quad core chip, without the FX branding, the disappointing would have being lower.

I'm still waiting AMD talk about a new revision to reduce the power use and increase the clocks, because the 10~15% performance increase they said that Piledriver will have, is very far from being enough to compete with Sandy Bridge, and Ivy Bridge won't get long to arrive.

If weren't by Llano and Bobcat, AMD would be in serious troubles, even with Bulldozers scoring good results on the server market.
__________________
Goodbye, and thanks for all the fish !
Cheetos is offline   Reply With Quote
Old 10-15-2011, 07:29 PM   #10
C.W.
Senior Member
 
C.W.'s Avatar
 
Join Date Jan 2011
Posts: 425
C.W. is on a distinguished road

Default

Those RAM tests are interesting as it adds a grain of salt to any review that uses 1333 RAM- yeah Bulldozer is no FX-51, but if you review the 8150 with just 1333, it's not showing the whole picture.


I agree that the new CPUs probably should have been hyped less and positioned as 32nm shrinks of the Phenom II line than something that can tackled the 2500K and 2600K SB processors. If they'd focused on trying to hit the sub $200 market (attack the 2400, 2300, and 2100) this would have worked. They could have marketed modules as advanced cores or something cool sounding. Then they could have gone Phenom III x2 (FX-4100), x3 (6100), and x4 (8120). This is sort of the strategy they came up with when it came to the 6xxx series of GPUs. The 6870 and 6850 from the beginning weren't marketed as crushing the 5870 and 5850. Instead, they set them up as the next level above the 5770 and 5750, but still within budget.

While the knee-jerk reaction is to call this an epic fail of a chip, I'd probably say it was an epic fail of marketing/sales.

Last edited by C.W.; 10-15-2011 at 07:33 PM.
C.W. is offline   Reply With Quote
new Sponsored Links

Reply

Share This Thread

Bookmarks

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Forum Jump


All times are GMT -8. The time now is 01:25 PM.


vBulletin Security provided by vBSecurity (Lite) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
vBulletin Optimisation provided by vB Optimise (Lite) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd. ()
© 2004-12, Hardware Secrets, LLC. All rights reserved.