Quote from Nick7 :

Clock speed is irrelevant for comparing CPU speeds between different generation/types of CPU's.

As an example is aforementioned P4 @3GHz compared to core2duo CPU at 1.8GHz (forgot which model it was).
Well.. P4 was SLOWER some 20-30%.

This is oversimplification. It doesn't matter if you have a CPU that can average 5 IPC in a certain application clocked at 3 GHz or a CPU averaging 10 IPC running at 1.5 GHz.
NetBurst wasn't a particularly efficient architecture, but the latest P4s would probably outperform the lowest-end C2Ds at certain kinds of workload. Since LFS doesn't take advantage of anything beyond SSE, I don't see why it would particularly benefit the Core2 architecture.

I have updated the LFS Benchmark site with the new version 0.6F and have added some hardware for those of us with new systems.

The replay and min/max cfg.txt files are still the same as with previous version.
It may be interesting to see if there's a performance difference between 0.6F with DirectX 9 vs the older versions of LFS with DirectX 8.

Instructions : http://lfsbench.iron.eu.org/?c=instructions
Max list : http://lfsbench.iron.eu.org/?c=completemax
Min list : http://lfsbench.iron.eu.org/?c=completemin

Thanks, added my results


Yet, again LFS seem to perform much better on Windows XP.



Also little comparison to old version 0.6E:

FPS is lower in newer version when using cfg with these settings.

However when running 1 car replay XFG+BL1 with all the settings maxed out, I have noticed increase of FPS in newer version.

Hm yes that's a big difference. But on the MIN list, the difference between your XP and Win7 results is minimal.
Don't have an explanation - just noticed it

I have noticed XP perform much better in variety of tasks.
For example filling up ~35000 items x 8 subitems ListView is atleast 2 times faster on XP. This has nothing to do with Aero or something. Also tried using different compilers from VC8 to VC12 same thing. Simple commands like send(), recv() take far less time to execute on XP...

Guess Scawen isn't that crazy after all

ummm, installed Windows 8.1 x64 to test FPS and only problems so far.


First tried running LFS with max settings, and I get only black screen on start, obviously something related to resolution, but however same resolution work on Windows 7 using absolutely same drivers (same exe on AMD site for both 7 and 8). My standard cfg work, as well as min cfg.


Tried min cfg benchmark then, and it runs, runs, almost reaching finish line, press button to stop benchmarking and LFS crashes:

Problem signature:
Problem Event Name: APPCRASH
Application Name: LFS.exe
Application Version: 0.0.0.0
Application Timestamp: 53a30b6e
Fault Module Name: apphelp.dll
Fault Module Version: 6.3.9600.16384
Fault Module Timestamp: 52158e39
Exception Code: c0000005
Exception Offset: 0006b7e2
OS Version: 6.3.9600.2.0.0.256.48
Locale ID: 1050
Additional Information 1: cd1d
Additional Information 2: cd1d258c1b4823adc6b7bda47ced280b
Additional Information 3: d5cb
Additional Information 4: d5cb7dbe23aaf1df8615e6b207df004b


And when I tried put report in CODE tags by clicking button found out it does not work. Windows 8.1 + IE umm neither smiles, umm nothing clickable


Looked up exports of DLL and this seems to be ReHook_CreateFileA(). In other words most likely FRAPS is causing LFS to crash.

Tried starting LFS "as administrator" and then crash does not happen :fp:

As nothing work on W8.1 and FPS is much lower I wouldn't recommoend anyone to switch (some wink smile or maybe facepalm)


and don't worry W8.1 is now on spare backup HDD, original HDD is untouched.

Grasshopper say, "The less resources an OS claims, the better it's results in any test which is single core focused."

If 7/8 are trimmed down then they'll perform as well as XP.

Quote from DANIEL-CRO :

ummm, installed Windows 8.1 x64 to test FPS and only problems so far.

On what ? I loath 8 but if you install DX9 in full then there seems to be no issue, if your running a lappy with intel graphics then I maybe understand but if your PC runs 8 then LFS should hardly raise a sweat.

I think I'm severely CPU limited.

The interesting thing for me is even if I run LFS @ 1024x960 or 2880x1800 the avg fps remains very similar

Here is my 2880x1800 results.

2014-07-05 19:20:57 - LFS
Frames: 8920 - Time: 108140ms - Avg: 82.485 - Min: 52 - Max: 110

Updated Linux benchmarking tool for 0.6F. The results are not entirely comparable with Fraps on Windows but hopefully the difference is minimal. Linux kernel 3.4 or above is required. Run

to build. See the "privileges.txt" file to set up advanced permission caps to avoid having to run this tool as root.

Gitweb link: http://gitweb.devoid-pointer.net/?p=LFSBench.git;a=summary

LOL my results here

Quote :

2014-07-05 17:10:07 - LFS
Frames: 14 - Time: 1172ms - Avg: 11.945 - Min: 12 - Max: 12

2014-07-05 17:10:09 - LFS
Frames: 660 - Time: 41875ms - Avg: 15.761 - Min: 1 - Max: 49

Quote from Litro :

LOL my results here

You didn't even run whole bench :bananadea

It's dead again?

Looks like the domain no longer exists. As it wasn't controlled by us, I can't bring it back.
Could make a new url though, if it's still interesting.

Think Always nice to have a place where u can test.

offcourse i don't know how much work it is