Hi, could anyone tell me what real world benefits there may be getting a dual core processor with 2x1Mb L2 cache as opposed to a similar clocked processor with 2x512Mb L2 cache?
I've had my eye on a 65w 4000+ windsor core (clocked @ 2.0 Ghz, 2x1Mb) that beats a 65w 4000+ brisbane core (clocked @ 2100GHz, 2x512Kb) on most benchmarks. Problem is, they're not so readily available anymore, and I'm wondering if I should really care about the 2x1Mb cache.
So I just spent ages comparing cpus and tallying things up on tom's cpu benchmark chart @ http://www23.tomshardware.com/cpu_2007.html
contenders were an e2140, e2160, 4000+ x2 windsor (2000 + 2x1), 4000+ x2 brisbane (2100 + 2x512), and 4200+ x2 brisbane (2200 + 2x512). they're all similarly priced @ around £35 to £40 on ebay, although the e2160 is the most expensice, as is the windsor 4000+ due to availability.
what I found was (and this is based on the number of benchmarks they were faster than the another cpu in, plus a bit of intelligence interpreting scores),
4200+ brisbane > e2160 (by a bit) and 4000+ windsor (kills it)
e2160 > 4000+ brisbane (not by too much)
4000+ windsor > 4000+ brisbane (by v v little, and real world progs) and e2140 (kills it)
4000+ brisbane > e2140
the fact the 1 Mb cache 4000+ beat the 512 Kb cache version by so little (basically, 1 bench mark between em, and 1Mb version won more real world prog tests) suggests there's very little difference in performance to gain by the larger cache, at least at those clock speeds. and when a slightly higher clock is involved, it takes presedence over cache amount and +200GHz gives a much better cpu even with -50% cache.
what i'm wondering is, am i missing anything? is there actually anything to gain for the larger cache size, and should i hold out to get the 4000+ windsor chip over either the brisbane version or 4200+ ?
oh and, this is ignoring all overclocking abilities. obviously the intels are gonna beat everything there, although... how oc'able is the 4200+ x2?
I've had my eye on a 65w 4000+ windsor core (clocked @ 2.0 Ghz, 2x1Mb) that beats a 65w 4000+ brisbane core (clocked @ 2100GHz, 2x512Kb) on most benchmarks. Problem is, they're not so readily available anymore, and I'm wondering if I should really care about the 2x1Mb cache.
So I just spent ages comparing cpus and tallying things up on tom's cpu benchmark chart @ http://www23.tomshardware.com/cpu_2007.html
contenders were an e2140, e2160, 4000+ x2 windsor (2000 + 2x1), 4000+ x2 brisbane (2100 + 2x512), and 4200+ x2 brisbane (2200 + 2x512). they're all similarly priced @ around £35 to £40 on ebay, although the e2160 is the most expensice, as is the windsor 4000+ due to availability.
what I found was (and this is based on the number of benchmarks they were faster than the another cpu in, plus a bit of intelligence interpreting scores),
4200+ brisbane > e2160 (by a bit) and 4000+ windsor (kills it)
e2160 > 4000+ brisbane (not by too much)
4000+ windsor > 4000+ brisbane (by v v little, and real world progs) and e2140 (kills it)
4000+ brisbane > e2140
the fact the 1 Mb cache 4000+ beat the 512 Kb cache version by so little (basically, 1 bench mark between em, and 1Mb version won more real world prog tests) suggests there's very little difference in performance to gain by the larger cache, at least at those clock speeds. and when a slightly higher clock is involved, it takes presedence over cache amount and +200GHz gives a much better cpu even with -50% cache.
what i'm wondering is, am i missing anything? is there actually anything to gain for the larger cache size, and should i hold out to get the 4000+ windsor chip over either the brisbane version or 4200+ ?
oh and, this is ignoring all overclocking abilities. obviously the intels are gonna beat everything there, although... how oc'able is the 4200+ x2?
i may convince the owner to let me up it a couple Ghz after a couple years, but i wouldn't wanna go much further than that anyway so it makes little difference what chip i use.
Facebook
Twitter