-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
Fritz Chess Benchmark/Super PI Thread 2010
- Thread starter easyrider
- Start date
Soldato
- Joined
- 20 Oct 2007
- Posts
- 2,778
- Location
- South Yorkshire
Dual Core vs Quad?
Why do the i3's perform so well at Pi, but so badly at Fritz?
Because super pi is a single thread application and therefore the result is mainly based on the frequency of the CPU and to some extent your RAM speed..
Intel's perform better than AMD on SuperPi because the program is deigned to use the Intel architecture when addressing the CPU cache
Associate
- Joined
- 19 Oct 2009
- Posts
- 169
- Location
- West Sussex
Just thought i'd up the game a bit.
I'm having real probs getting Fritz to run at anything higher than above.
It could be a BCLK issue, although i can run Prime and Linx fine at 234 BCLK.
I will try again to get it to run over 5ghz, but not today.
this is as high as i can get stable at my settings 
no 10k for me
q6600 @ 3.5
PI - 14.927
FRITZ - 9359
close but no cigar
should put me in the middle of a phenom sandwich though
and thanks paradigm realtemp seems to work fine
i have no idea why cpu-z locks me up
i'm pretty sure i'd get over 10k if i ran it on xp but i wouldn't feel right doing that
no 10k for me 
q6600 @ 3.5
PI - 14.927
FRITZ - 9359
close but no cigar
should put me in the middle of a phenom sandwich though
and thanks paradigm realtemp seems to work fine
i have no idea why cpu-z locks me upi'm pretty sure i'd get over 10k if i ran it on xp but i wouldn't feel right doing that
Last edited:
This is running:
AM3 945 (95w) @ 3.6ghz
8gb PC6400 OCZ Platinum RAM
Asus M3NHT-Deluxe (AM2+)
I couldn't get the chess program to run any higher than 3.6ghz without crashing my system but hey i'm not complaining coming from a chip with a locked Multi
edit: I have had it running at 3.85ghz seemingly stable running Prime95 but the Fritz chess always seemed to cause a BSoD at these settings
Last edited:
OK, at last update there were 47 entries in the table, probably pushing mid 50s with the latest posts.
Now, I know I took it on myself to put the results in a table, nobody asked me etc, but do we really need everyones' results? 20 odd 4ghz i7s all scoring within 20 points of each other helps nobody get a quick overview of the performance differences between CPUs.
Basically what I'm asking is for suggestions of ways to thin out the results to a more meaningful sample, rather than just listing the results of everyone who happens to stumble into this thread.
Ideas so far are:
Just keep the top 50 - Good for epeen, terrible for getting an idea of performance of different chips.
Fastest clocked of each CPU - Keeps a range of chips, but may give an unrealistic idea of what kind of OC to expect.
Stock speeds only - Terrible idea for an enthusiast forum, also doesn't give any indication of expected performance. Very few people here run any chip at stock.
Fastest and lowest clocked of each CPU - Potentailly OK but there's one issue. Where do you draw the line of lowest clock? Stock, Mild OC etc.
Mean results from each cpu - This will find average score and average OC, but will make updating the table a terrible job as I'll need to redo the whole thing from scratch every update, rather than just adding new data.
Only one entry per person - good for limiting table size, bad if people want to compare the same chip at different speeds.
Thoughts? Better suggestion?
At this point I'm unsure what direction the thread is taking, best scores or wide spread of CPUs tested. Obviously how I thin out the results will pretty much shape this decision, hence the need for imput.
easyrider, it's your thread so your call will be final.
Everyone else, I know I just asked for opinions but we don't want the thread cluttering anymore than needs be. Only respond to this if you have a very different idea to those I suggested, or you feel very stongly for/against one of the suggestions.
Now, I know I took it on myself to put the results in a table, nobody asked me etc, but do we really need everyones' results? 20 odd 4ghz i7s all scoring within 20 points of each other helps nobody get a quick overview of the performance differences between CPUs.
Basically what I'm asking is for suggestions of ways to thin out the results to a more meaningful sample, rather than just listing the results of everyone who happens to stumble into this thread.
Ideas so far are:
Just keep the top 50 - Good for epeen, terrible for getting an idea of performance of different chips.
Fastest clocked of each CPU - Keeps a range of chips, but may give an unrealistic idea of what kind of OC to expect.
Stock speeds only - Terrible idea for an enthusiast forum, also doesn't give any indication of expected performance. Very few people here run any chip at stock.
Fastest and lowest clocked of each CPU - Potentailly OK but there's one issue. Where do you draw the line of lowest clock? Stock, Mild OC etc.
Mean results from each cpu - This will find average score and average OC, but will make updating the table a terrible job as I'll need to redo the whole thing from scratch every update, rather than just adding new data.
Only one entry per person - good for limiting table size, bad if people want to compare the same chip at different speeds.
Thoughts? Better suggestion?
At this point I'm unsure what direction the thread is taking, best scores or wide spread of CPUs tested. Obviously how I thin out the results will pretty much shape this decision, hence the need for imput.
easyrider, it's your thread so your call will be final.
Everyone else, I know I just asked for opinions but we don't want the thread cluttering anymore than needs be. Only respond to this if you have a very different idea to those I suggested, or you feel very stongly for/against one of the suggestions.
Associate
- Joined
- 15 Jun 2009
- Posts
- 595
2nd place? I'll take that
Last edited:
One entry, per CPU, per person. So if for example I benched my Phenom II x3 720BE, I could have that in the table too.
There should be enough differentiation between people's overclocks to get a good judge of different clock speed with each type of CPU to negate the need for multiple entries per person with the same chip.
That way, everyone gets their fastest run on the table, and they aren't limited to a single architecture.
It's a modified Johann SS phase unit.
Soldato
- Joined
- 23 Nov 2009
- Posts
- 6,032
- Location
- North Leicestershire
my idea would be keep 3 overclocks from each processor the best stock fritz attemp the mid range attemp and the high range attemp
example
my phenom 955 is currently leading at 4.1Ghz this would be the high range overclock attemp the mid range would be 3.6/3.8 and stock would obviously be 3.2 this should keep the list down quite a bit as their would only be 3 scores for each cpu and at a glance would be able to make some sort of relative speed/cost/performance info.
think i7's should be stock,3.5 and 4.2GHZ
i'm sure with a bit of colaboration we could come up with a idea that best suits everyone.
example
my phenom 955 is currently leading at 4.1Ghz this would be the high range overclock attemp the mid range would be 3.6/3.8 and stock would obviously be 3.2 this should keep the list down quite a bit as their would only be 3 scores for each cpu and at a glance would be able to make some sort of relative speed/cost/performance info.
think i7's should be stock,3.5 and 4.2GHZ
i'm sure with a bit of colaboration we could come up with a idea that best suits everyone.
highest entry, per person, per cpu seems fair, here are mine in list form for ease of entry
q6600 @ 3.5 abit ip35 pro
PI - 14.927
FRITZ - 9359
core duo t2300e @ 1.66 hewlett packard 30a6 intel i945pm
PI - 36.015
FRITZ - 2167
core2duo e4300 @ 1.8 some dell p.o.s
PI - 31.824
FRITZ - 2447
athlon xp 1600 @ 1.4 pc chips M863G (V1.5A)
PI - 86.391
FRITZ - 656
athlon 64 3500+ @ 2.2 msi amethyst ati express 200 skt 939 rs480 chipset i think
PI - 42.407
FRITZ - 1264
celeron e3200 @ 2.4 gigabyte GA-G41M-ES2L
PI - 24.648
FRITZ - 3275
q6600 @ 3.5 abit ip35 pro
PI - 14.927
FRITZ - 9359
core duo t2300e @ 1.66 hewlett packard 30a6 intel i945pm
PI - 36.015
FRITZ - 2167
core2duo e4300 @ 1.8 some dell p.o.s
PI - 31.824
FRITZ - 2447
athlon xp 1600 @ 1.4 pc chips M863G (V1.5A)
PI - 86.391
FRITZ - 656
athlon 64 3500+ @ 2.2 msi amethyst ati express 200 skt 939 rs480 chipset i think
PI - 42.407
FRITZ - 1264
celeron e3200 @ 2.4 gigabyte GA-G41M-ES2L
PI - 24.648
FRITZ - 3275
Last edited:
Seems perfectly logical to me.
It also gives peeps a good incentive to get as good a clock as possible out of whatever rig they are running. After all, this is an Overclocking forum
