VOGONS


Battle of the platforms: socket 939!

Topic actions

Reply 580 of 600, by nd22

User metadata
Rank l33t
Rank
l33t

9. 3dmark 2001 1600*1200:
A. Athlon 64 X2 3800
B. Athlon 64 X2 4200
C. Athlon 64 X2 4400
D. Athlon 64 X2 4800

Reply 581 of 600, by nd22

User metadata
Rank l33t
Rank
l33t

And just when you were thinking things are going AMD's way we got this: scaling crumbles as we increase the resolution and we go back to Aquamark3 scores where 4800 is performing more like a 4400-4500 and the lower models perform way too close for their rating!

Reply 582 of 600, by nd22

User metadata
Rank l33t
Rank
l33t

15. I misplaced the screenshots for the Athlon 64 X2 4400 in 3dmark 2003 so until I find them let's skip 2003 and go to 3dmark 2005!
A. Athlon 64 X2 3800
B. Athlon 64 X2 4200
C. Athlon 64 X2 4400
D. Athlon 64 X2 4800

Reply 583 of 600, by nd22

User metadata
Rank l33t
Rank
l33t

4800 is more like a 4600; 4400 perform like a 4300 and the 4200 shows the performance of a hypothetical 4100. Again scaling is poor!

16. 3dmark 2005 1280*1024 max details;
A. Athlon 64 X2 3800
B. Athlon 64 X2 4200
C. Athlon 64 X2 4400
D. Athlon 64 X2 4800

Reply 584 of 600, by nd22

User metadata
Rank l33t
Rank
l33t

What do you know? The tougher settings make for an absolute perfect scaling! Every single model has the score its PR implies it should have!

17. 3dmark 2005 1600*1200 max details:
A. Athlon 64 X2 3800
B. Athlon 64 X2 4200
C. Athlon 64 X2 4400
D. Athlon 64 X2 4800

Reply 585 of 600, by nd22

User metadata
Rank l33t
Rank
l33t

One more time AMD performance rating is spot on! Incredible isn't it?

18. The last version of 3dmark compatible with Windows XP is the 2006 one. The resolution went up from 1024*768 to 1280*1024 to better reflect the 17&19 inches LCD's that invaded the market in the first wave during the transition from CRT to LCD.
A. Athlon 64 X2 3800
B. Athlon 64 X2 4200
C. Athlon 64 X2 4400
D. Athlon 64 X2 4800

Reply 586 of 600, by nd22

User metadata
Rank l33t
Rank
l33t

19. Resolution stays at 1280*1024 but the details are all maxed out this time around:
A. Athlon 64 X2 3800
B. Athlon 64 X2 4200
C. Athlon 64 X2 4400
D. Athlon 64 X2 4800

Reply 587 of 600, by nd22

User metadata
Rank l33t
Rank
l33t

20. And finally 1600*1200; even at this high resolution and settings GTX 285 is still limited by all Athlon 64X2 models!
A. Athlon 64 X2 3800
B. Athlon 64 X2 4200
C. Athlon 64 X2 4400
D. Athlon 64 X2 4800

Reply 588 of 600, by nd22

User metadata
Rank l33t
Rank
l33t

All 4 models tested scale far below the AMD ratings: 4800 like a 4500, 4400 like 4300 and 4200 like a 4100! So far, with few exceptions, AMD PR does not stand up to scrutiny!

Reply 589 of 600, by nd22

User metadata
Rank l33t
Rank
l33t

10. We go back to 3dmark 2003; I retested this morning because I was missing the screenshots.
Default settings first:
A. Athlon 64 X2 3800
B. Athlon 64 X2 4200
C. Athlon 64 X2 4400
D. Athlon 64 X2 4800

Reply 590 of 600, by nd22

User metadata
Rank l33t
Rank
l33t

Results of the processors with 512kb of L2 cache are grouped together just like the results of the processors with 1mb of L2 cache!

11. 1280*1024 max details
A. Athlon 64 X2 3800
B. Athlon 64 X2 4200
C. Athlon 64 X2 4400
D. Athlon 64 X2 4800

Reply 591 of 600, by nd22

User metadata
Rank l33t
Rank
l33t

Less than 9% separate the strongest -X2 4800 - from the weakest CPU - 3800!

12. 3dmark 2003 1600*1200
A. Athlon 64 X2 3800
B. Athlon 64 X2 4200
C. Athlon 64 X2 4400
D. Athlon 64 X2 4800

Reply 592 of 600, by nd22

User metadata
Rank l33t
Rank
l33t

All results are within 7% - practically no scaling at all!
13&14. 3dmark 2003 is the only version with 3 sound tests: no sound and 24 sounds that can be run on any audio card including the crappy AC97 included in nforce4 and 60 sounds for which you need a hardware accelerated solution such as Creative Audigy! Between the weakest - X2 3800 - and the strongest - X2 4800 - there is only 10% difference!
Conclusion for 3dmark 2003: poor scaling makes the Performance rating of X2 unsustainable!

Reply 593 of 600, by nd22

User metadata
Rank l33t
Rank
l33t

21&22. CINEBENCH R10 is next; despite the fact that multiprocessor usage is not 200% but only 180% something, this is very good actually!
A. Athlon 64 X2 3800
B. Athlon 64 X2 4200
C. Athlon 64 X2 4400
D. Athlon 64 X2 4800

Reply 594 of 600, by nd22

User metadata
Rank l33t
Rank
l33t

Again the results swings back in AMD favor! The differences is just about where the performance ratings indicate it should be for both single thread and multi threads scores!

23. Next is Wprime benchmark which knows how to make use of dual cores, quad cores and even more!
A. Athlon 64 X2 3800
B. Athlon 64 X2 4200
C. Athlon 64 X2 4400
D. Athlon 64 X2 4800

Reply 595 of 600, by nd22

User metadata
Rank l33t
Rank
l33t

Please note that a shorter time constitutes a better result here! Just 15% separates the top of the line 4800 from the lowly 3800! And the results swing against AMD!

24. 7-ZIP, version 9.20 from 2010, is the next test and again it knows how to use multiple threads!
A. Athlon 64 X2 3800
B. Athlon 64 X2 4200
C. Athlon 64 X2 4400
D. Athlon 64 X2 4800

Reply 596 of 600, by nd22

User metadata
Rank l33t
Rank
l33t

All it matters is the frequency here: 4400 behaves almost like a 4200! Again higher clocked Athlon 64 X2 do not deserve their rating!
Next is the 2 PCMARK versions that test general system performance: 2004 and 2005. I left out the 2002 version because it only knows how to use single cores! PCMARK 2004 "knows"about 2 cores/threads, while PCMARK 2005 "knows" how to use 4 cores/threads.

Reply 597 of 600, by nd22

User metadata
Rank l33t
Rank
l33t

25 & 26 & 27. PCMARK 2004 tests are completely overhauled from the ones found in the 2002 version; it shows you a general score and individual scores for your CPU, RAM, GPU and HDD/SSD which are the one we want. The last score (storage) is not actually taken into account and it will not appear in the final table!

Reply 598 of 600, by nd22

User metadata
Rank l33t
Rank
l33t

Disaster struck as the amount of level 2 cache simply does not matter: the CPU and memory scores for 4200 and 4400 are practically identical! Also what is far worse is the graphics score that is within 3% across all 4 models of Athlon 64 X2 tested! All that matters is the frequency: 200 MHz more leads to 10% better CPU score; 400 MHz more leads to 20% better score.
I am still debating about including the graphics score in the final Excel table! Maybe PCMARK 2005 will show more differentiation between the X2?

Reply 599 of 600, by nd22

User metadata
Rank l33t
Rank
l33t

28 & 29 & 30. PCMARK 2005 is the very last version for Windows XP. An enhanced version of the 2004 one, it has more tests suited for multi-threaded CPU's.
A. Athlon 64 X2 3800
B. Athlon 64 X2 4200
C. Athlon 64 X2 4400
D. Athlon 64 X2 4800