You're right I suppose. Looking at this comparison table http://www.tomshardware.com/ch ... re,818.html?prod=on∏=on the C2D seems to be about 40-50% faster in games but not double like with LFS. However they've used much greater resolution in their gaming tests so the GPU might be more of a limiting factor there, dunno.
You're right I suppose. Looking at this comparison table http://www.tomshardware.com/ch ... re,818.html?prod=on∏=on the C2D seems to be about 40-50% faster in games but not double like with LFS. However they've used much greater resolution in their gaming tests so the GPU might be more of a limiting factor there, dunno.
SuperPi has always been better on Intel, even when other benchmarks show the systems to be the same.
(For reference - overclocked K8 3700+ at 3 GHz - 26 seconds. My Quad at stock 2.13 GHz - 16 seconds!)
(For reference - overclocked K8 3700+ at 3 GHz - 26 seconds. My Quad at stock 2.13 GHz - 16 seconds!)
Sometimes it seems to work and sometimes not. Can't really fix it right now but you could simply keep trying in the mean time.
lets not talk about superpi 
From memory there is only a handful of games that display such a tendency to run upto 40% faster at comparable clock cycles (Intel/AMD) and LFS is one if not the game that displays this level of advantage to owning an Intel System. (damn beat me to it vari)
Does make me wonder if i could replace the CPU string for my system, would it show any anomolies
Is LFS a cache fiend ? It Would be interesting to see some low end core2 with castrated amounts of L2
About the registration problems Vari, last night i updated my score. But at first i could not remember my password, so i went to the forgot/reset page.
Entered my email and got a similar if not the same error as Toki.
After closing down browser and starting again it seemed to work
From memory there is only a handful of games that display such a tendency to run upto 40% faster at comparable clock cycles (Intel/AMD) and LFS is one if not the game that displays this level of advantage to owning an Intel System. (damn beat me to it vari)
Does make me wonder if i could replace the CPU string for my system, would it show any anomolies
Is LFS a cache fiend ? It Would be interesting to see some low end core2 with castrated amounts of L2
About the registration problems Vari, last night i updated my score. But at first i could not remember my password, so i went to the forgot/reset page.
Entered my email and got a similar if not the same error as Toki.
After closing down browser and starting again it seemed to work
Jakg -Intel has NOT always been better at super pi. Athlons 64s would beat Intel P4s clock for clock and to compare an intel core2 quad against an althlon xp clock for clock is utterly stupid - the intel does load more operations per clock cycle.
As I said, Althon 64s used to own P4s in LFS.
Gerer_AL - what's your pi time? As Vari said, most games are gpu bound and so don't show the true difference between cpus. LFS and super pi are both cpu tests and as such it would be interesting to compare.
As I said, Althon 64s used to own P4s in LFS.
Gerer_AL - what's your pi time? As Vari said, most games are gpu bound and so don't show the true difference between cpus. LFS and super pi are both cpu tests and as such it would be interesting to compare.
I didn't mean P4 vs K8, I meant recently - a lower clocked Core 2 versus a top-of-the-line AMD CPU which wouldn't be that far apart in benchmarks would be a lot different in SuperPi.
When did I mention an Athlon XP? I was just saying that my old CPU, a 3700+ at 3.07 GHz (which isn't really much slower than a 5600+ or whatever it's called, besides being single core) got 26 seconds...
When did I mention an Athlon XP? I was just saying that my old CPU, a 3700+ at 3.07 GHz (which isn't really much slower than a 5600+ or whatever it's called, besides being single core) got 26 seconds...
ok sorry, I thought 3700+ was an althon xp, it's an althon 64.
Anyway, the core2 still does more operations per clockcyle than the a64 or phenom, that's why it's quicker at the same or less clockspeed.
Compare the 3700+, ~3ghz vs my 3.4ghz P4 extreme (2mb cache) - 26seconds vs 36 seconds again because the A64 did more operations per clock cycle than the P4.
I really think that you'll see that LFS and super pi show very similar results when comparing cpus.
Anyway, the core2 still does more operations per clockcyle than the a64 or phenom, that's why it's quicker at the same or less clockspeed.
Compare the 3700+, ~3ghz vs my 3.4ghz P4 extreme (2mb cache) - 26seconds vs 36 seconds again because the A64 did more operations per clock cycle than the P4.
I really think that you'll see that LFS and super pi show very similar results when comparing cpus.
The problem should be solved now. Thanks V
Maybe someone here knows the technical reasons for the Intel chips being such good performers in LFS and SuperPi? I'd like to know, and I'd like to know, is it more the Intel doing very well, or the AMD doing badly, or a combination of both maybe..
I'm getting 28 seconds in SuperPi Athlon64 6400+BE stock speeds, with 667Mhz memory. I'm assuming 800Mhz memory will bring it under 25, and when running @ 3.5 should give me a decent score, I dunno. I start building my new PC soon, I guess I'd still like to know though, also it's been a long time....I hope AMD has something planned, soon, they need a new mainstream chip, and not 90nm!!! I'm using twice the power and making half the speed of a good intel chip, uhg.
I'm getting 28 seconds in SuperPi Athlon64 6400+BE stock speeds, with 667Mhz memory. I'm assuming 800Mhz memory will bring it under 25, and when running @ 3.5 should give me a decent score, I dunno. I start building my new PC soon, I guess I'd still like to know though, also it's been a long time....I hope AMD has something planned, soon, they need a new mainstream chip, and not 90nm!!! I'm using twice the power and making half the speed of a good intel chip, uhg.
Basically, Intel had the P4 where they thought clock speed pwned.
AMD realised that being the smaller company they would have to try to be different to win - so they went to better architecture, which performed better at a lower clock speed - they refined this with the Athlon XP and then to the 754/939 chips. They kicked Intel's ass for years - Intel decided that they couldn't afford for this to happen and made Core 2, which runs even more stuff per clock cycle, runs cooler, uses less power etc - since then they've been "keeping AMD on the ropes" and so AMD just can't compete.
Basically - Intel just did better.
To give you an idea of how much things have changed - Core 2 can do SuperPi in half the time that a P4 can (when both are overclocked). Your top-of-the-range 6400BE Dual Core does SuperPi 2 seconds faster than my 3 year old single core 3700+ when overclocked - you even have an extra 500 MHz under your belt, and DDR2 when I had DDR1...
AMD realised that being the smaller company they would have to try to be different to win - so they went to better architecture, which performed better at a lower clock speed - they refined this with the Athlon XP and then to the 754/939 chips. They kicked Intel's ass for years - Intel decided that they couldn't afford for this to happen and made Core 2, which runs even more stuff per clock cycle, runs cooler, uses less power etc - since then they've been "keeping AMD on the ropes" and so AMD just can't compete.
Basically - Intel just did better.
To give you an idea of how much things have changed - Core 2 can do SuperPi in half the time that a P4 can (when both are overclocked). Your top-of-the-range 6400BE Dual Core does SuperPi 2 seconds faster than my 3 year old single core 3700+ when overclocked - you even have an extra 500 MHz under your belt, and DDR2 when I had DDR1...
I've made a poll about the benchmark here: http://www.lfsforum.net/showthread.php?t=48957
You have to realize that when you say "always" you extend the timeline quite a bit. Plus, P4 has always been shit, no matter how you turn it. C2D is another story all together.
The site has been updated with a new replay
Major changes:
-1280*960 resolution
-4XAA/8XAF for the 'max' benchmark (min is 0X/0X)
-Full GTR Grid in SO4R (full S2 needed)
I lost 60 fps, happy benchmarking
http://lfsbench.iron.eu.org/
Major changes:
-1280*960 resolution
-4XAA/8XAF for the 'max' benchmark (min is 0X/0X)
-Full GTR Grid in SO4R (full S2 needed)
I lost 60 fps, happy benchmarking
http://lfsbench.iron.eu.org/
Bump ,
I see that a Z28 replay has been added.
Meh i lost quite abit of frames there but the old benchmark was done with 8xAA/16xAF when the new one is done with 4xAA/8xAF which just makes it worse .
Sidenote: the AMD Phenom X3 series is not in the CPU list but the X2/X4/X6 are.
,I see that a Z28 replay has been added.
2008-10-01 01:24:53 - LFS
Frames: 7525 - Time: 86628ms - Avg: 86.865 - Min: 58 - Max: 1292010-10-07 14:20:53 - LFS
Frames: 27303 - Time: 108623ms - Avg: 73.686 - Min: 33 - Max: 120Meh i lost quite abit of frames there but the old benchmark was done with 8xAA/16xAF when the new one is done with 4xAA/8xAF which just makes it worse .
Sidenote: the AMD Phenom X3 series is not in the CPU list but the X2/X4/X6 are.
I've added the X3
I've just added my benchmarks. Oh god, I wish my new PC arrived 
Added my benchmark too.
added...
Who runs this thing?
My benchmark has been deleted from the top (user boola)!! I will re-upload it later...
My benchmark has been deleted from the top (user boola)!! I will re-upload it later...
Is anyone who runs this able to update the replays to 0.6B? The old replay gets OOS in one part.
done 
a bit of hardware question, I have E5200@4GHz and think it should be more than enough for LFS, but for some reason when on cargame with 25+ cars (mostly same cars) I got like 3-5 FPS
. CPU usage is 55-60 % (coz of dual core), and GPU barely hits 15%. Why cargame races eating so much my CPU? Same thing applies for just watching replay of actual race. Sound is crackling...
a bit of hardware question, I have E5200@4GHz and think it should be more than enough for LFS, but for some reason when on cargame with 25+ cars (mostly same cars) I got like 3-5 FPS
. CPU usage is 55-60 % (coz of dual core), and GPU barely hits 15%. Why cargame races eating so much my CPU? Same thing applies for just watching replay of actual race. Sound is crackling...Got mine up too
Is there any reason to use Fraps 1.9 (2.1 version is from 2004), or I could use a modern version?
i think you can use more modern versions of fraps. The 1.9 one is linked to because it was still a free version. But if your modern fraps produces the same frame info logging output, then that's fine as well.
LFS Benchmark
(139 posts, started )
FGED GREDG RDFGDR GSFDG