• 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.

Cache may affect gaming

AGD

AGD

AGD

Soldato
Joined
23 Nov 2007
Posts
5,048
I was quite interested in whether or not the differences in cache between the intel core 2 duos would make a difference in games bearing in mind that the e2180 can be oc'd so if cache doesn't make a difference then why bother with anything else (quads aside)?

Lots of remarks here seem to suggest that l2 cache doesn't really affect modern game performance but having seen these 2 tests I'm not so sure.

http://www.nordichardware.com/Reviews/?page=9&skrivelse=514

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2795&p=4

The first one shows considerable difference between 1mb and 4mb cache.

The second compares only 2mb and 4mb and whilst in most cases there is no appreciable difference you still get a not insubstantial 5.4% difference in an oblivion dungeon. The difference between 4mb and 1mb would probably be greater.

I'm not an expert on the validity of these tests and know that overall the e2180 is much better value but was just wondering if it is misleading to say the difference is negligible in modern games.
 
It is confusing because two seemingly respectful sources come to different conclusions. It seems that because most games are GPU bound and atm that cache is not the most important factor. It would be nice to see some really new games tested though like bioshock or crysis. Maybe if someone was intending to upgrade their gpu when the 9 series comes out then it would be worth it...unless they bought a penryn at the same time. That's slightly too many factors though.
 
alexthecheese said:
That is an excellent summary of cache, FSB and cores for stuff that we can actually relate to. Cache = good. FSB = not to worry about. Cores = good. Ghz = not fussed. Cheers. :cool:
Click to expand...

Well, only for Conroes: obviously if you increase the clockspeed without increasing the cache proportionately you'll see less benefit, as higher clockspeeds need the CPU to be fed more data. (AMDs seem to be less reliant on cache due to their lower memory latency.) Conversely, if you double the cache without increasing hte clock speed you'll achieve very little speed increase, as the CPU's pipelines will always be backed-up waiting for the current bit of number crunching to finish before its ultra-fast cache can feed it the next bit.
 
I thought someone on here ran some bench marks recently of an e21XX vs a e6600 or something similar, and the differentce was at times like literally maximum of a few %, but at other times the e21XX was amazingly actually every so slightly better (both clocked the same).
 
Yeah, it's Melbourne720's thread that I linked to above, but he's using older games, whereas anandtech are testing with UT3 which is brand new and uses the CPU more. However, anandtech are testing at a fairly low resolution so as to take the stress away from the graphics card completely and focus solely on the CPU. As you up the resolution, the E2180's disadvantage will become smaller as the CPU matters less and less and the GPU more and more. At 1280x1024 I'm sure those bar charts will look a lot flatter and that 20% difference will decrease to 5-10%. At 1600x1280 they'd probably look pretty much like Melbourne720's graphs.
 
I've read in a couple of places that for Athlon X2s increasing the cache per core from 512k to 1MB is equivalent to about a 100MHz increase in clock speed.
 
You must log in or register to reply here.
Back
Top Bottom