Abit KR7A-raid with VIA KT266A chipset is next:
1024*768; 1280*1024; 1600*1200

Far Cry Benchmark

The benchmark started at 09.10.2024 20:35:36

System Information
Operating system: Microsoft Windows XP
System memory: 2,0 GB
CPU: AMD Athlon(tm) processor
CPU speed: 1200 MHz
Sound system: : SB Live! Wave Device
VGA Information
Graphics card: NVIDIA GeForce2 MX/MX 400
Memory: 256.0 MB
Current GPU speed: 560 MHz
Current memory speed: 700 MHz
Driver version: 6.14.0010.9371 (English)

Resolution: 1024×768
Ultra quality option, Direct3D renderer
Level: Volcano, demo: hocvolcano.tmd
Pixel shader: default model
Antialising: 8×
Anisotropic filtering: 16×
HDR: disabled
Geometry Instancing: disabled
Normal-maps compression: disabled

Score = 31,80 FPS (Run 1)
Score = 31,19 FPS (Run 2)
Score = 31,53 FPS (Run 3)
Average score = 31,50 FPS

Resolution: 1280×1024
Ultra quality option, Direct3D renderer
Level: Volcano, demo: hocvolcano.tmd
Pixel shader: default model
Antialising: 8×
Anisotropic filtering: 16×
HDR: disabled
Geometry Instancing: disabled
Normal-maps compression: disabled

Score = 32,06 FPS (Run 1)
Score = 31,53 FPS (Run 2)
Score = 31,93 FPS (Run 3)
Average score = 31,83 FPS

Resolution: 1600×1200
Ultra quality option, Direct3D renderer
Level: Volcano, demo: hocvolcano.tmd
Pixel shader: default model
Antialising: 8×
Anisotropic filtering: 16×
HDR: disabled
Geometry Instancing: disabled
Normal-maps compression: disabled

Score = 31,84 FPS (Run 1)
Score = 31,07 FPS (Run 2)
Score = 31,60 FPS (Run 3)
Average score = 31,50 FPS

Copyright 2002 - 2006 Zoltan Nemeth - Roadside

Abit AN7 with nforce2 ultra chipset:

Far Cry Benchmark

The benchmark started at 06.10.2024 16:46:27

System Information
Operating system: Microsoft Windows XP
System memory: 2,0 GB
CPU: AMD Athlon(tm) processor
CPU speed: 1200 MHz
Sound system: : NVIDIA(R) nForce(TM) Audio
VGA Information
Graphics card: NVIDIA GeForce 7600 GT
Memory: 256.0 MB
Current GPU speed: 560 MHz
Current memory speed: 700 MHz
Driver version: 6.14.0010.9371 (English)

Resolution: 1024×768
Ultra quality option, Direct3D renderer
Level: Volcano, demo: hocvolcano.tmd
Pixel shader: default model
Antialising: 8×
Anisotropic filtering: 16×
HDR: disabled
Geometry Instancing: disabled
Normal-maps compression: disabled

Score = 34,66 FPS (Run 1)
Score = 33,71 FPS (Run 2)
Score = 34,24 FPS (Run 3)
Average score = 34,20 FPS

Resolution: 1280×1024
Ultra quality option, Direct3D renderer
Level: Volcano, demo: hocvolcano.tmd
Pixel shader: default model
Antialising: 8×
Anisotropic filtering: 16×
HDR: disabled
Geometry Instancing: disabled
Normal-maps compression: disabled

Score = 34,79 FPS (Run 1)
Score = 34,06 FPS (Run 2)
Score = 34,54 FPS (Run 3)
Average score = 34,46 FPS

Resolution: 1600×1200
Ultra quality option, Direct3D renderer
Level: Volcano, demo: hocvolcano.tmd
Pixel shader: default model
Antialising: 8×
Anisotropic filtering: 16×
HDR: disabled
Geometry Instancing: disabled
Normal-maps compression: disabled

Score = 34,46 FPS (Run 1)
Score = 33,48 FPS (Run 2)
Score = 33,84 FPS (Run 3)
Average score = 33,92 FPS

Copyright 2002 - 2006 Zoltan Nemeth - Roadside

Farcry summary: nforce2 is the undisputed leader in this game since I started testing! nothing can touch NVidia top chipset for socket A!
On the same time SDRAM and KT133A exhibit a ever worsening result as the resolution is increasing!

41&42&43 The second game is the legendary Half life 2! This is the original 2004 version and not the Steam version. I tested right in front of the train station where Dr Breen is making his speech using max details as the geforce 7600gt is more than capable of handling this game.
I used Fraps 1.9D because HardwareOC bench tool refused to work. I am sorry for the lack of screenshots - Fraps 1.9D outputs the results to a .TXT file.

First is KT7A-raid:
1024*768
2024-10-12 05:21:07 - hl2
Frames: 1361 - Time: 74026ms - Avg: 18.385 - Min: 9 - Max: 26

1280*1024
2024-10-12 05:19:17 - hl2
Frames: 1375 - Time: 76259ms - Avg: 18.030 - Min: 6 - Max: 28

1600*1200
2024-10-12 05:17:35 - hl2
Frames: 1037 - Time: 68188ms - Avg: 15.207 - Min: 9 - Max: 23

Second is Abit KR7A-raid

1024*768
2024-10-09 17:55:55 - hl2
Frames: 1478 - Time: 62069ms - Avg: 23.812 - Min: 7 - Max: 43

1280*1024
2024-10-09 17:53:56 - hl2
Frames: 1369 - Time: 61178ms - Avg: 22.377 - Min: 17 - Max: 29

1600*1200
2024-10-09 17:52:19 - hl2
Frames: 1219 - Time: 62009ms - Avg: 19.658 - Min: 13 - Max: 26

Third is Abit AN7

1024*768
2024-10-06 17:36:14 - hl2
Frames: 1554 - Time: 60219ms - Avg: 25.805 - Min: 9 - Max: 36

1280*1024
2024-10-06 17:34:51 - hl2
Frames: 1343 - Time: 54453ms - Avg: 24.663 - Min: 16 - Max: 36

1600*1200
2024-10-06 17:33:34 - hl2
Frames: 1228 - Time: 55141ms - Avg: 22.270 - Min: 4 - Max: 30

Half life 2 is not actually playable at max details on any of the systems simply because the 1200B is killing the performance. 7600Gt can run just fine the game at max details including AA however it is bottlenecked in the extreme by the processor.
Looking at the average framerates KT133A is far behind the rest, while nforce2 ultra provides the best results and wins across all 3 resolutions in this game.

47. Doom 3 with RoE can not run on Thunderbird because it requires SSE so we go to Super Pi where a shorter time to calculate 1 million decimals is a better result!
KT7A-RAID
KR7A-RAID
AN7

SDRAM cripples again the KT133A platform and it is 24% slower than its DDR brother, the KT266A.
Nforce2 ultra scores the last victory in this comparative - 8% better than KT266A!

Cinebench 2003 should be the last test however it requires SSE so no results for this one!

The final battle, at 200 MHz FSB is over! Time for the rankings!
KT266A is awarded a 100% and the other 2 platforms are compared to it:
1. ABIT AN7 nforce2 ultra: 108.30%
2. ABIT KR7A-RAID Via KT266A: 100%
3. Abit KT7A-RAID Via KT133A: 81.11%%

For the last time nforce2 ultra wins! Despite not being shown here I tested the other boards and the closest to nforce2 performance is Abit KV7/VIA KT600, still 5% behind the leader!
VIA chipsets KT333; KT400 have indistinguishable performance from KT266A! Nforce1 is not worth mentioning; AMD760 is always behind KT266A.

Now for some things that can not be seen in the screenshots:
-on every single DDR platform Windows XP felt snappy and fluid; there was no slowdowns, startup and shutdown were speedy consindering hard disks were used in every single one!
-on both SDRAM system XP takes ages to start and shut down; loading Farcry 1 at max details took enough time that i was able to make me a hot chocolate 😀 as I was switching resolutions!

Now that all tests are finished we can compare the 2 Athlons running at 1200 MHz, one with a 200 MHz FSB and the other one with a 266 Mhz FSB.
Based on the results whose screenshots can be seen on the previous pages the Athlon 1200C is 10.33% faster than 1200B! Impressive advantage for what is only a change in FSB!

I will now try to answer some possible questions:
- first and probably the one that will spring to everyone's mind: but period correct reviews such as those done by Anandtech - https://www.anandtech.com/show/687/12- and Xbitlabs - https://web.archive.org/web/20041104003400/ht … pox-8kta3.html- showed no more then 5% difference between Athlons with the same frequency but different FSB! How can the difference be double in your tests?
The video card I used no longer constitutes a bottleneck; in period correct reviews a geforce2 GTS/ultra was used which is actually a big bottleneck for the Athlon; period correct review also did additional testing in content creation which I did not as I am not interested.
- another question related to the first one: how come the difference between SDRAM chipsets - VIA KT133A for example - and DDR chipsets - AMD760 for example - is so small in period correct reviews but you found a staggering 14% difference between KT133A and AMD760, not to mention the 17% advantage KT266A has over KT133A?!?
My selection of tests has a very heavy focus on 3d performance because I believe that socket 462 main purpose is and will remain retro gaming; furthermore I tested at different resolution and level of details and in all cases as the resolution and LOD grew so does the advantage of DDR over SDR grew!
- your selection of games is poor; many synthetics tests and very few games!
Guilty as charged; I selected 3 DirectX 9.0 games only, I tried using NFS underground 1&2 but it was impossible to obtain repeatable results; The Chronicles of Riddick: Escape from Butcher Bay is another game that I tried to use but could not get reliable results; in fact I believe that including 3dmark 2005 in the test suite was a mistake; despite being released in 2004 it is actually very tough on the video card! I left out DirectX 8 and 7 games because they are too easy to run on most socket A systems and I wanted to push the systems to the limits.

I made a document where I put all the results for both the 1200B and 1200C on nforce2 ultra - the advantage of 10% is on nforce2 as it is the highest performing platform:

Having completed a comparison between 200 MHz and 266 MHz processors I decided to also compare the difference between 266 MHz and 333 MHz ones. The main testing will remain focused on 3d performance again!
Unlike Thunderbirds where AMD released multiple models with the same frequency but different bus between 1000 MHz all the way to 1400 MHz, in the case of CPU's with a 266 and 333 MHz FSB there are very few models with the same frequency, cache and only FSB different! In fact we have to look at the Sempron line of CPU to find equally clocked models!

Sempron are in fact based on 2 different Athlon XP revisions: most of them - from 2200 to 2800 models - are based on the Thoroughbred B core; in fact Sempron is a rebadged Thoroughbred in order to keep the Athlon name reserved for the expensive models such as Athlon 64. There is architecturally no difference between Sempron and Thoroughbred B with a single exception: FSB! They all have a 333 Mhz FSB which is perfect! There are also Semprons based on the Thorton core which is in fact a Barton CPU with some defective cache deactivated and are from all points of view identical to the Semprons based on the Thoroughbred-B core.
There is one model - Sempron 2400 - which has the same clock speed and cache as the Thoroughbred B 2000. The only difference is the advantage in FSB that the Sempron has over Thoroughbred.
There is also the Sempron 2800 which has an identical clock speed and cache with the Thoroughbred B 2400. I do not have that model yet so all testing will be done on the Sempron 2400.

There is also a Sempron model based on the full Barton core, no cache deactivated: Sempron 3000. The Sempron 3300 is for all intents and purposes a Athlon XP 3200 with a different name - again AMD tried to keep the Athlon name reserved for the "noble"models!

All testing will be done on Abit AN7/nforce2 ultra motherboard. Having won every single competition with every Athlon classsic/XP generation it is the best performing motherboard for socket462. There is no need to test again different models with different chipsets, the testing is focused on the difference in performance between models with the same clock speed and cache but different FSB.
All previous used components will remain the same:
GPU: Leadtek geforce 7800gs
RAM: 2*1gb from Corsair as these are stable with the same 2-2-2-5 timings at both 266 MHz and 333 Mhz FSB.
HDD: WD raptor 150gb - the fastest SATA1 drive ever manufactured!
ODD: some PATA model I have laying around
OS: Windows XP SP3 with no updates
Cooler: some Spire model with a cooper core

Nice comparisons!
It's so funny how things change. Back in the day, I desperately wanted an nForce 2 Ultra 400 platform, but nowadays... I couldn't care less about it. The KT133A is much more impressive and capable to me.

Thank you very much.
Contrary to you my favorite retro system for the last 9 years, the only one that remains permanently fully assembled, is the nforce2 one: Abit AN7 with 2gb of RAM and geforce 7800gs plus Athlon XP 3200 -that CPU changes as I need to run the tests on different models - runs every single game from 2000 up to 2004 perfectly. In fact all of my gaming is done on this system! It runs incredible well XP and all games that I have installed, even games made for Windows9X such as Undying - I am talking about original release not GOG release!