Main > Main Forum

4Ghz Core 2 Duo vs M.A.M.E. 0.120 (benchmark results)

<< < (98/104) > >>

Diet_Pepsi:
It would be interesting to see if you needed all 4.8ghz to run these games or if they are able to run at, say, 4.2 or 4.3.

smalltownguy:
Yeah, that's an interesting point to consider. Since I was overclocking for the pure purpose to see how high I could push my chip and still remain stable, I'm not inclined to mess with my settings now that I've got it locked in stable at 4.8.

I would, however, be happy to bench test any other games that folks would like to see benchmarked.

Diet_Pepsi:

--- Quote from: smalltownguy on February 28, 2012, 04:07:40 pm ---Considering I can see MAME64 running in two of my eight threads @ 4.8ghz each, I can see where a single core CPU would need to run close that speed in order to keep up.



--- End quote ---

I am curious -- while your benchmark is running how many of your threads are at 100% -- 2 or 4?  If only 2 of 8 are at 100% then shouldn't you be able to disable HT to get a higher benchmark (mame would then use 2 of 4)?  I'm going to try this out on my Clarkdale when I get home...for some reason I can't shake the notion that HT negatively affects mame performance.

smalltownguy:

--- Quote from: Diet_Pepsi on February 28, 2012, 04:28:11 pm ---
--- Quote from: smalltownguy on February 28, 2012, 04:07:40 pm ---Considering I can see MAME64 running in two of my eight threads @ 4.8ghz each, I can see where a single core CPU would need to run close that speed in order to keep up.



--- End quote ---

I am curious -- while your benchmark is running how many of your threads are at 100% -- 2 or 4?  If only 2 of 8 are at 100% then shouldn't you be able to disable HT to get a higher benchmark (mame would then use 2 of 4)?  I'm going to try this out on my Clarkdale when I get home...for some reason I can't shake the notion that HT negatively affects mame performance.

--- End quote ---

My motherboard has a lot to do with how well my i7 performs. It does a very good job with the voltage during TurboBoost. I had CPU-Z running during my benching, and the board did a very good job ratcheting up the core speeds from 3.4 (stock) to 4.8ghz when under load. AFAIK the turbo boost mode overclocks all 8 cores, so they're all running at 4.8 regardless of whether or not a process is running in that core or not.

I believe MAME64 v145 only uses 2 cores at most. I could be wrong. I'll shoot urebelscum a PM and ask.

Diet_Pepsi:
OK so I was able to spend some time tinkering.  Got my i3-560 up to 4.2ghz (going higher would require more time than I was willing to invest) and ran some benchmarks with and without HT.  I am running mame64 in windows 7 (ultimate 64 bit) with 8 gb of RAM.  I had a limited selection of roms but I was able to benchmark the following:

galaga (no HT): 5415
galaga (HT): 3973

1943 (no HT): 4124
1943 (HT): 4114

1942 (no HT): 5175
1942 (HT): 5209

gauntleg (no HT): 275
gauntleg (HT): 263

So it looks like 1942 and 1943 are pushes, galaga runs significantly faster (?!) and gauntleg has about a 4.5% advantage with no HT.  I took a closer look while running the gauntleg bench 90 and my CPU has a slightly higher overall utilization when HT is turned off.

These limited results don't indicate that turning off HT gives a significant advantage but they do indicate that turning on HT doesn't benefit mame64.  I will therefore be leaving HT off (which will lower temps and allow for a higher overclock).

Out of curiosity I fired up gauntleg and it ran just fine until it crashed while I was on the stairs (known issue).  I didn't have blitz handy but I will pull it out tomorrow and post results.

Navigation

[0] Message Index

[#] Next page

[*] Previous page

Go to full version