shevalier wrote on 2025-10-01, 12:36:

Mobile Athlon 64 4000+ will surpass everyone Or, at worst, Turion ML-44 They are rare and impossible to obtain. […]
Show full quote

AlexZ wrote on 2025-10-01, 12:02:

The expectation is Athlon 64 in s754 should be able to squeeze more out of late AGP cards than Athlon XP. There are also AGP boards for s939, but those are rare whereas in s754 they are very common and cheap. The downside is availability of high clocked CPUs like 3200+/3400+ is not good. s939 is superior, but s754 could be "good enough". The main advantage of s754 is very low price, nobody wants them, they are cheaper than Athlon XP systems but slightly more modern.

Mobile Athlon 64 4000+ will surpass everyone
Or, at worst, Turion ML-44
They are rare and impossible to obtain.

2.4-2,6GHz + 1Mb l2 cache
And high multiplier will not limit FSB (In fact, Hypertransport) overclocking.

i have a 4000+ Mobile Athlon but it doesnt run on NF6G-VSTA Mainboard

AlexZ wrote on 2026-06-17, 19:34:

Company of Heroes (2006) benchmark with Turion 64 at 2400Mhz and GTX 275:

The game needs to be patched to 1.4.0 for results to be comparable. It has a built-in benchmark.

It runs very well with maximum settings.

I know this game very well; it is one of my favorites and I still play it! What version of the game have you used? It must be one of the earliest version because AA is shown only as enabled! I played just now a skirmish with the game fully patched up and both expansions installed: opposing fronts and tales of valor. Furthermore the in-game test does not shows the actual performance during game play; it is more of a video card test!
These are screenshots with the options and the version of the game: 2.700.2.42 is the latest version.

Company of heroes under maximum details is extremely demanding! The game with both expansions installed and fully patched can not be played on any single core CPU! Even the first dual cores do not provide a fluid frame rate! No socket 939 CPU, not even Athlon 64 X2 4800 can not play this game satisfactory!
In addition it requires a video card far beyond everything available in 2005-2006; with AA = 4X and ultra details GTX 280 is the minimum! Also 4gb of RAM are mandatory: it will take around 2.5-2.6gb of RAM and that is under Windows XP.

This is a game that I started just now at 1024*768, ultra details, AA = 4x. please note that the 512mb frame-buffer of the 9800gtx is already full and the 2.3gb (!!!) of ram - and this is not the largest map available. Company of heroes is wiping the floor with both Athlon 64 3500 and geforce 9800GTX with averages in the low 20's and minimums in the single digits!! I am starting to believe that dual cores are needed far sooner than universally accepted! more testing to be done!

I used Company of Heroes (2006) patched to 1.4.0. It is vanilla, without any expansion packs. The installer scared me with warning about hardware demands, but the game runs really well. You only get into a bit of trouble at very low camera angle when trying to view too much area. The version you use is probably newer and more demanding. I observed the same with other games.

The next game is FEAR Perseus Mandate (2007), version 1.0. It falls into Windows Vista era (2007-2009).

Frame rate is a bit lower than FEAR First Encounter Assault Recon (2005), but not much.

Maximum details, FSAA 4x, Anisotropic 16x, SS off. Enabled HW audio mixing and EAX 2.0 since we have Audigy 2 ZS. SS off because it always resets after exiting the game.

Since the Company of Heroes (2006) benchmark can be misleading I started skirmish for 4 players with 3 AIs - Lorraine map.

See screenshots for different camera angles. Unless you set very low angle the game runs fine. Normally you would play probably with angle like in the 2nd screenshot with about 40 fps. Game resolution is 1280x1024.

I think you really need GTX 275/280/285 with 896 MB VRAM

Prey (2006) benchmark with Turion 64 at 2400Mhz and GTX 275:

The game needs to be patched to 1.4.0. Benchmarked using Prey Benchmark tool.

We get comfortable 100+ fps.

Splinter Cell Chaos Theory (2005) benchmark with Turion 64 at 2400Mhz and GTX 275:

The game needs to be patched to 1.0.5. Benchmarked using Splinter Cell Chaos Theory Benchmark tool.

We get comfortable 130+ fps. Very easy for s754 to handle.

matze79 wrote on 2026-06-17, 20:41:

shevalier wrote on 2025-10-01, 12:36:

Mobile Athlon 64 4000+ will surpass everyone Or, at worst, Turion ML-44 They are rare and impossible to obtain. […]
Show full quote

AlexZ wrote on 2025-10-01, 12:02:

The expectation is Athlon 64 in s754 should be able to squeeze more out of late AGP cards than Athlon XP. There are also AGP boards for s939, but those are rare whereas in s754 they are very common and cheap. The downside is availability of high clocked CPUs like 3200+/3400+ is not good. s939 is superior, but s754 could be "good enough". The main advantage of s754 is very low price, nobody wants them, they are cheaper than Athlon XP systems but slightly more modern.

Mobile Athlon 64 4000+ will surpass everyone
Or, at worst, Turion ML-44
They are rare and impossible to obtain.

2.4-2,6GHz + 1Mb l2 cache
And high multiplier will not limit FSB (In fact, Hypertransport) overclocking.

i have a 4000+ Mobile Athlon but it doesnt run on NF6G-VSTA Mainboard

The BIOS on the 754 socket is rather strangely designed.
In it, the initial CPU initialisation—if there is no separate microcode for it—is carried out by the main BIOS body.
There is no microcode for Venice and mobile CPUs.

For example, my Jetway with an Award BIOS and a Athlon Venice E6 core behaves as follows
- BIOS ‘last -2’ – unknown Hammer Athlon
- BIOS ‘last -1’ – Sempron
- BIOS ‘last ’ – Athlon 3000+

It seems that AMI takes a stricter approach to this compared to Award BIOS.

Maybe some of you who have E6 cpu can test it in various boards, although it should be just Semprons.

I found E5 CPU (Lancaster - ML/MT) to work with both AWARD and AMI BIOS. MSI K8T Neo V (VIA K8T800 chipset) has AMI BIOS and CPU is detected correctly.

Mobile Athlon 64 4000+ (Newark) is E5 as well.

If you are referring to desktop models I already tested Venice 3200 in Abit KV8 PRO in the other topic: Re: Battle of the platforms: socket 939!
I do not have a Sempron and I will not get any as Athlon 64 for socket 754 are so easy to source: just the other day I got for free an Athlon 64 3200 clawhammer CO stepping - the original Athlon released back in September 2003.
If you are referring to mobile models - those cost money and i already got 9 desktop models so I see no point in spending extra.

AlexZ wrote on 2026-06-19, 09:33:

Maybe some of you who have E6 cpu can test it in various boards, although it should be just Semprons.

I found E5 CPU (Lancaster - ML/MT) to work with both AWARD and AMI BIOS.

Mobile Athlon 64 4000+ (Newark) is E5 as well.

No, it’s a full-fledged Athlon DH-E6 3000+ with 512kb L2 cache.
One of its highlights is its very low power consumption (51Watt vs 89Watt Athlon 64 3000 clawhammer C0).
And the marking that includes ‘BX’
That’s why I always say that ‘the best BX is the Athlon BX’ 😀

https://www.cpu-world.com/CPUs/K8/AMD-A ... XBOX).html

https://pctuning.cz/article/116x-firmware-a-b … e-40-a-41-tyden
K8T8AS BIOS version A11
Remove AMD logo for feature,AMD K8 REV:E6 support , RAID BOOT ROM update V4.90

PS
If anyone has a TRW account, please upload the BIOS to the downloads section.
https://theretroweb.com/motherboards/s/jetway-k8t8as
The longer it goes on, the harder it is to find.
At the moment, it can only be found via WebArchive.
Jetway K8T8AS A11 bios file
https://web.archive.org/web/20060624005853/ht … as/K8TAIA11.BIN

Unreal Tournament 3 Black Edition (2009, base game from 2007) benchmark with Turion 64 at 2400Mhz and GTX 275:

Game version is 2.1. Benchmarked using UT3 Benchmark Tool (version 0.2.0.45). HOC Unreal Tournament 3 Bench 1.3 crashes the game.

Once you play Unreal Tournament 3, you will not look back at Unreal Tournament 2004. It's so superior in every way.

This game is quite tough for s754, being from Vista era. The engine is better optimized than Crysis and delivers playable experience but on the low end. I consider 50 fps on average ok, but not great experience. We get maybe 46 fps across all maps. You want at least 60+ average fps for good experience. This game could be a good benchmark for s939 to differentiate single and dual core.

s754 didn't do that well in Quake 4 benchmark, just 44 average fps but in actual game it runs much better, with about 60 fps (engine limit).

Results for CFT Coret map and DM Sactuary. Used 1280x1024 and 1600x1200. Resolution has little effect as we are CPU bottlenecked. Because these are bot matches, the result is not repeatable but is very accurate when it comes to game experience.

Results for 1600x1200 in various maps. Those with ~43 fps are open maps, very tough on the system. Those with ~50 fps are closed maps.

The game is playable even on open maps, however you would be disadvantaged due to latency.

This is just to show limits of s754. You would like to play historical FPS games at 60+ fps across all maps, not 43 fps.

Company of Heroes Complete Edition (2009) benchmark with Turion 64 at 2400Mhz and GTX 275:

There was an objection that Company of Heroes (2006) is not the latest, one would like to enjoy the game with expansion packs so it would make sense to retest it on the latest version to see if it can be played on s754.

4x AA was used, the rest is maximum quality. Some options allow Ultra quality that cannot be selected (Shader Quality, Terrain Detail). It can probably be enabled only with dx10. Environmental reverb cannot be enabled.

We see a major drop in built-in benchmark over the 1.4.0 version. Minimum in 1280x1024 is probably just a glitch, it looks ok in 1600x1200. Since average fps goes down noticeably, we are no longer GPU limited. We would benefit from GTX 285 or even GTX 480.

Results from skirmish game, Lorraine map with AI in 1280x1024. Same map was used for 1.4 version, although we are in different position.

Camera angle 2 would typically be used in the game and we get even better results than in 1.4, although due to different position we cannot make conclusion that the game runs better. It did very well for the chosen camera angle. 1 GB swap file on magnetic hard drive was enabled not to run out of memory. All apps were closed to conserve memory as we have just 2 GB RAM.

We have sufficient FPS headroom for combat.

Based on performance, I would say the game is playable and enjoyable on s754 with a powerful GeForce GTX 275/280/285 GPU.

Let's try a map for 6 players, just like nd22 did:

Initially we start with decent 66 fps, but it drops to 33 fps once we get to part of map with many buildings. Later, once enemy invades with tanks it drops to 18 fps. Basically same behavior as Age of Empires 3. The game consumes about 660MB VRAM. A GPU with at least 768 MB VRAM is highly recommended for s754.

While not a slideshow, it is clear on large maps, with lot of units we would benefit from X2 CPU. On s754, you should probably play maps for max 4 players.

AlexZ wrote on 2026-06-18, 19:00:

Prey (2006) benchmark with Turion 64 at 2400Mhz and GTX 275:

The game needs to be patched to 1.4.0. Benchmarked using Prey Benchmark tool.

We get comfortable 100+ fps.

Where did you get the Prey and Splinter cell benchmark tools from? I suppose they should be pretty small, maybe you can upload them.

I got both of them from guru3d.

https://www.guru3d.com/download/prey-benchmark-1/
https://www.guru3d.com/download/splinter-cell … eory-benchmark/

Those are 20 years old pages so get them.