Intel vs AMD Gaming CPU Showdown 2500K vs 1100T with Quad SLI Linus Tech Tips

Linus Tech Tips ·Linus Tech Tips ·2012-05-07 · 1,463 words · ~7 min read
Floatplane YouTube

Transcript

JSON SRT VTT 105
0:07 AMD finally has a high-end a truly
0:11 high-end platform this so this is the 990fx Crosshair 5 formula with an 1100t
0:19 8 gigs of RAM a 120 gig SSD it's my
0:22 usual test platform except it has two
0:26 GTX 590s on it so this is pretty much
0:29 the the fastest gaming machine you can
0:33 build on an AMD
0:36 CPU well 6990 would be pretty much equivalent so so it'd be an equivalent
0:41 fastest gaming machine so what I'm going to be doing is I'm going to be looking
0:45 at once and for all which is better for
0:49 gaming AMD or Intel CPUs at least from
0:53 this generation which won't be relevant uh very soon with bulldozzer coming out
0:57 and Intel having a refresh coming but whatever C so we'll know for now so I'm
1:01 going to do the 1100t versus the Intel
1:04 Core i 52500 k now the reason that I'm
1:08 using these two particular chips is that
1:11 they are fairly equivalently priced at this time so there is no uh price to
1:16 Performance ratio argument to be made really I mean the 2500k has to win by
1:21 about 10% just to break even so uh yeah
1:25 let's see how this goes this is going to be mostly Graphics Bound in spite of the
1:29 fact fact that I'm running quad SLI with 590s cuz I am going to be running fairly
1:34 intensive games I may add some more games to the test Suite but let's just
1:38 see how things go and uh this will be interesting I've uh I've wondered this
1:42 for a long time how do these two platforms compare against each other in
1:46 a truly equivalent scenario oh yeah for the Intel uh CPU I will be using the
1:52 z68a gd80 as my motherboard of choice so
1:56 I've loaded on the Intel platform now this is a Z6 68 platform but I will not
2:02 be using the onboard uh
2:05 Graphics at all so I'm going to be disabling it because turning on virtue
2:10 actually will impact graphics performance somewhat while it will
2:15 increase video encoding speed through supported program so I'm just going for
2:19 peer graphics performance we're assuming this is for a gaming system given that
2:23 there are dual GTX 590s in it so I've finished my 1100t results and I will be
2:28 following up now with my 2500k results I wanted to give you guys a little bit of
2:33 info first though so for all of these I'm running at 2560
2:37 x600 and I'm going with my usual settings so that means for example
2:42 Witcher 2 I'm not using the very highest setting I'm using High uh for example uh
2:47 Battlefield Bad Company 2 Forex antialiasing 8times AF uh Fear 3 I had
2:52 to turn off AA even though normally I use it at 2560 x600 I was getting only
2:57 about 12 FPS which is not enough for me to play through my run through with the
3:01 1100t so I'm turning off antialiasing that made a big difference to
3:04 Performance so we'll see how that goes crisis 2 I am using the
3:10 um shoot I can't remember for the life of me oh yeah the the second the second
3:14 highest that's right and then Dirt 3 I have it on Ultra so here we go 2500k
3:20 testing upcoming quad SLI is working let's see how this goes one other thing
3:24 to mention actually Mafia 2 I am using Apex physics on high 8X AF and anti-al
3:30 lasing is set to on so I'm using NVIDIA
3:33 cards might as well crank up the physx all right so I have my results with the
3:37 2500k this is all using the same graphical settings so in Dirt 3 we see a
3:43 12.5% increase in performance so this is about what I would expect given the CPU
3:48 costs about 10 to 12% more okay in
3:51 crisis 2 we see a 35.5% difference in performance I mean
3:56 62.5 frames per second is substantially more playable than 46.8 just due to the
4:01 fact that you're not going to see those same dips um Witcher two we saw just a
4:06 very insubstantial increase in performance so we're obviously not very
4:10 CPU bound here uh fear three now this one's funny I'm going to go back I'm
4:14 going to put the 1100t with the 990 FX back on the test bench and I'm going to
4:19 run it again just to be sure but I did run into some weirdness on the AMD
4:24 platform with fear 3 where when I was running
4:27 anti-aliasing uh the performance was about a third of this it was very very
4:31 low and just turning off AA brought us up to here whereas on the Intel platform
4:37 with with zero times anti-aliasing I didn't see any any weirdness in terms of
4:41 performance Fear 3 is not a super demanding game so even at 2560 x600 I
4:46 was expecting more like this number and that's what I got so I'll rerun it just
4:49 to just to see Battlefield Bad Company 2 it's about par for the course and then
4:54 Mafia 2 uh maybe due to the processor
4:57 intensive nature of it with all the articles and whatnot and the CPU still
5:01 has to help out the GPU with that so we ended up seeing a 40 approximately 40%
5:07 increase in gaming performance going to a 2500k from an 1100t with GTX
5:15 590s so let's um you know what I'm actually going to since I have to go
5:18 throw the other platform back on anyway what maybe what I'll do is I'll run uh
5:22 I'll run one more game here just cuz it's not quite super super late yet so I
5:27 can do a little bit more testing for you guys so thinking maybe I'll do um Lost
5:32 Planet 2 yeah let's do Lost Planet 2 so I made myself a smoothie and still at
5:37 work here so I ran Metro 2033 and Lost
5:41 Planet 2 so Metro 2033 actually showed
5:44 the 2500k losing this is the first time I've seen that and then Lost Planet 2
5:49 showed a 20% victory for the 2500k now
5:53 Fear 3 I I put that back on Fear 3 has
5:56 um exhibited exactly the same behavior second time around so the 2500k even
6:01 with four times anti-aliasing still manages a better frame rate than fear 3
6:05 on the 1100t without anti-aliasing so thinking I mean SLI support is pretty
6:11 new for AMD platform so maybe there's an issue with the SLI driver what I think
6:15 I'm going to do well not I think I'm going to do what I'm going to do right now is I'm going to take a 68 oh 68 I
6:21 don't even have a 68 well I do have a 6870 but I don't have it on me so
6:27 6970 there we go so I'm run 69 70s In
6:31 Fear 3 alone with the same graphic settings 0 x anti-aliasing a single card
6:36 I don't think it's going to run that well anyway and then I'm going to run that on both platforms and I'm going to
6:39 see if it's CPU related or if it's
6:43 related to the quad SLI driver all right
6:46 well it seems to be a weird driver issue in retrospect maybe I should have chosen
6:50 a more normal Graphics configuration like maybe uh GTX 580 SLI or something
6:56 like that I couldn't use 6970 Crossfire because I know crisis 2 behaves very
7:01 strangely with it from some other benchmarking I did this weekend anyway
7:04 you can see here the 1100t with the 6970 gets 86.2
7:10 FPS uh uh at that setting at 2560 x600
7:14 with this yeah so so that's a substantially less powerful
7:17 configuration it has like about one qu the GPU processing power so now we're
7:22 going to take the 2500k run it with the 6970 see how it comes out and that is
7:27 what we will take to be our performance difference for fear
7:31 3 well that was rather anticlimactic after all that fear 3 is a dead tie so
7:37 thank you for checking out my coverage of the 2500 K versus the 1100t in Quad
7:44 SLI and In Fear 3 running a 6970 don't
7:48 forget to subscribe to Linus Tech tips for more unboxings reviews and other
7:52 computer videos