21. The final 3dmark is the 2003 version. Again we start with default settings:
A. Athlon XP 3200
B. Athlon 64 3200

22. Athlon 64 is 7% better at default settings, maybe we got a GPU bottleneck!
1280*1024
A. Athlon XP 3200
B. Athlon 64 3200

23. Actually there is no video card bottleneck - the distance between the 2 increases to 10% as we up in resolution!
1600*1200
A. Athlon XP 3200
B. Athlon 64 3200

In 3dmark 2003 Athlon 64 scores better as the resolution is increased - XP 3200 is strongly bottlenecking the geforce 7800gs!
24 & 25. 3dmark 2003 is the only version with a sound test included; only sound storm can do the 60 sounds test so it is not taken into account.
Most people expect nforce2 to win over K8T800 thanks to sound storm! However the results show the opposite: Athlon 64 with VIA destroys the Athlon XP with nforce2 with a 50% (!!!) lead in the NO SOUND test and 36 % (!!) in the 24 sounds test! It seems that the famous nforce2 can not help Athlon XP at all as its bigger cousin walks again all over it!

26 & 27 & 28. Pcmark 2002 is the first benchmark that test all around system performance. Please note that identical Raptors are used in both systems.

A. Athlon XP 3200
B. Athlon 64 3200

Athlon 64 scores only 5% higher than XP which it is not such a big deal; however..
Dual channel? What dual channel? Athlon 64 got a 39% lead over XP in the memory tests which proves that having an integrated memory controller is far more important than having dual channel on the motherboard!
Also note that I tested multiple memory configurations on socket A - 3*512mb, 2*1gb, 3*1gb and the memory score remains the same - and yes, Abit AN7 "stays"in dual channel with 3 memory modules:

29 & 30 & 31 & 32. Pcmark2004 which adds a few graphics tests is next:
A. Athlon XP 3200
B. Athlon 64 3200

Please note how close are the graphics and hard drive scores - practically the same! So Silicon image 3112 performs the same as the VT8237 south bridge.
Athlon 64 platform scores 12% better in the CPU department and a staggering 36% better in the memory department! It should be clear by now that the Athlon 64 is the victor, the only question is by how much?

33 & 34 & 35 & 36 PCMARK 2005 is the final version:
A. Athlon XP 3200
B. Athlon 64 3200

One more time the IMC of the K8 enables the Athlon 64 to secure a 29% lead in the memory department; there is a 7% difference between the K8/K7 in the CPU score and again the graphics and storage results are extremely close!

37. We finally reached THE GAMES! The first one is Unreal tournament 2004, one of the last DirectX8.1 games ever made. Because Umark software refused to run I used Fraps 3.4.7 to benchmark a 60 seconds interval on the death-match map Antalus! This was used by many review sites back in the day. After each resolution change I waited 60 seconds before pressing F11 again!

1024*768

A. ATHLON XP 3200
Frames Time (ms) Min Max Avg
6136 60000 68 161 102.267

B. ATHLON 64 3200
Frames Time (ms) Min Max Avg
8291 60000 94 220 138.183

38. At 1280*1024 we get the following results:

A. ATHLON XP 3200
Frames: 5932 - Time: 60000ms - Min: 63 Max: 160 Avg: 98.867

B. ATHLON 64 3200
2025-11-22 06:23:25 - UT2004
Frames: 6984 - Time: 60000ms - Avg: 116.400 - Min: 72 - Max: 210

39. And finally at 1600*1200:

A. ATHLON XP 3200
Frames: 5415 Time: 60000ms Min: 59 Max: 135 Avg: 90.25

B. ATHLON 64 3200
2025-11-22 06:22:03 - UT2004
Frames: 6707 - Time: 60000ms - Avg: 111.783 - Min: 83 - Max: 164

Athlon 64 simply demolishes Athlon XP! Even at 1600*1200 there is a 23% lead - simply outstanding performance by the new micro-architecture!

40 & 41 & 42. FARCRY is next; very tough on the CPU, I tested at max details as the 7800gs is more than capable of handling it!

A. Athlon XP 3200
B. Athlon 64 3200

25% advantage at 1024*768 is enormous; it shrinks to just 8% at 1600*1200 - still significant!

43. Half life 2 is next. This is the original version which is far easier on the video card than the current steam version! All settings at maximum.

1024*768
A. Athlon XP 3200
Frames: 3807 - Time: 60000ms - Min: 50 - Max: 85 - Avg: 63.45

B. Athlon 64 3200
2025-11-22 05:16:28 - hl2
Frames: 4730 - Time: 60000ms - Avg: 78.833 - Min: 58 - Max: 98

44. As you already guessed I resorted to Fraps because HardwareOC half life 2 benchmark tool refused to work.

1280*1024

A. ATHLON XP 3200
Frames: 3732 - Time: 60000ms - Min: 51 - Max: 77 - Avg: 62.2

B. ATHLON 64 3200
2025-11-22 05:13:42 - hl2
Frames: 4437 - Time: 60000ms - Avg: 73.950 - Min: 60 - Max: 96

45. 1600*1200

A. ATHLON XP 3200
Frames: 3314 - Time: 60000ms - Min: 36 - Max: 72 - Avg: 55.233

B. ATHLON 64 3200
2025-11-22 05:12:16 - hl2
Frames: 4118 - Time: 60000ms - Avg: 68.633 - Min: 49 - Max: 90

Again a clear victory for the Athlon 64!

49. We reached the final game: Doom 3. Because this one is very GPU demanding I tested with AA disabled in order to not overwhelm the poor 7800gs:

1024*768

A. Athlon XP 3200
B. Athlon 64 3200

This game perfectly illustrates the limitation of both machines; even with AA disabled there is such a strong video card limitation that the Athlon 64 manages to best the Athlon XP at this low resolution by only 10% despite winning convincingly until now!

50. At 1280*1024 there is no more difference between the 2:

A. Athlon XP 3200
B. Athlon 64 3200