Intel i3 vs AMD Phenom 2

yeah psu first i would say, if you put in the new MB, CPU and RAM and its too much for your PSU it could fail and take out your new components, play it safe and get the psu first. Also get a trusted brand PSU so that you are safe knowing that its not likely to fail :)
 
Moothead2 said:
Stulid, most of the games in the link you provided are around 2+ years old and will be unlikely to use more than 2 cores effectively. I honestly think though that OP should keep their current rig or whatever and save up the extra 30-50 quid and get an i5k setup.

I did a bit of searching and I found a 2500k setup for around 270 quid.
Click to expand...

The i3 2100 is in this months Custom PC mag (wont be to long till the review is up at Bit-tech) and guess what. They compared it to the AMD 980BE and it(i3 2100) got better results in Gimp image editing tests, the H.264 video editing tests, Crysis etc.
 
Last edited:
FrenchTart said:
That's another suite of tests that use low settings for the games. Utterly pointless.
Click to expand...
Em...no. Lower resolution means it minimising things like GPU limitation, and comparing CPU performance directly. Your reasoning on applies if they were comparing graphic cards performance, but not for test that are using the same graphic card.
 
Marine-RX179 said:
Em...no. Lower resolution means it minimising things like GPU limitation, and comparing CPU performance directly. Your reasoning on applies if were comparing graphic card performance.
Click to expand...

No. This is the exact thing I'm talking about. Comparing the performance of a CPU in a game in a situation that would not occur in real life is absolutely pointless.

Results of games running at super low res might show one CPU hitting 200 FPS and the other only hitting 150 FPS. Oh noes! CPU2 is 25% slower than CPU1! Of course, once you run games in real settings and real situations, the actual difference between the two CPUs is probably in the region of 1-3 FPS, equivalent of a single digit percentage of performance difference.
 
FrenchTart said:
No. This is the exact thing I'm talking about. Comparing the performance of a CPU in a game in a situation that would not occur in real life is absolutely pointless.

Results of games running at super low res might show one CPU hitting 200 FPS and the other only hitting 150 FPS. Oh noes! CPU2 is 25% slower than CPU1! Of course, once you run games in real settings and real situations, the actual difference between the two CPUs is probably in the region of 1-3 FPS, equivalent of a single digit percentage of performance difference.
Click to expand...
You see any frame rate there in those modern games being in 3 digits as you claim? What that benchmark is showing is what frame rate each CPU is capable of delivering, when on the equal ground of the same graphic card . Also, Metro2033 is probably the one game that used low res for comparison, due to the graphic card can't handle the higher res.

And also, extra 5fps on minimum and on average is A LOT for most people. Hell, just look at those people that making a fuss and condemning 1GB graphic card, when 2GB is only deliver around 3fps higher in Metro2033 and 0fps difference in games in general.

The point is...going higher res won't make a CPU that's slower suddenly becomes faster another CPU that's faster than it at lower res. And with Phenom II X4 roughly consuming twice the power comparing to the i3 2100 and only able to deliver similar performance, slower by quite a margin in games that making use of all four of its coures, plus needing to spend extra on getting CPU cooler on keeping the temp under control...it simply doesn't make sense to build a new system base on Phenom II X4 over the i3 2100 for gaming and general usage, let alone the faster SandyBridge CPUs.
 
Last edited:
Marine-RX179 said:
You see any frame rate there in those modern games being in 3 digits as you claim? What that benchmark is showing is what frame rate each CPU is capable of delivering, when on the equal ground of the same graphic card .

And also, extra 5fps on minimum and on average is A LOT for most people. Hell, just look at those people that making a fuss and condemning 1GB graphic card, when 2GB is only deliver around 3fps higher in Metro2033 and 0fps difference in games in general.
Click to expand...

The numbers were examples used to explain the discrepancy between the two testing methodologies. I would've hoped that would be obvious... :rolleyes:

You're also still missing the point I'm making. Results for non-real world situations can not be applied to real world situations. (underlined it for you as I know you like that)

The minimum frames issue you've brought up is a completely separate one. Of course minimum frame rates matter. The frame rate differences I mentioned were not specified as being minimum frame rates...in fact, given the high numbers quoted in the first part of my post it should have been clear that the frame rates were likely to be the maximum frame rates or at least the average frame rates.
 
Last edited:
Marine-RX179 said:
The point is...going higher res won't make a CPU that's slower suddenly becomes faster another CPU that's faster than it at lower res. And with Phenom II X4 at 3.7GHz consuming twice the power comparing to the i3 2100 to deliver similar, plus needing to spend extra on getting CPU cooler on keeping the temp under control, slower by quite a margin in games that making use of all four of its coures, there's really nothing makes the Phenom II X4 a better buy than the i3 2100, let alone the faster SandyBridge CPUs.
Click to expand...

You seem to be missing the point. I'll beat my head against the brick wall one more time though I guess.

Graph of the PIIx4 vs the i3
should_be_obvious.png

And this is ignoring games where the extra cores of the X4 make a difference (e.g. BFBC2, etc)

edit: You can hit a reasonable OC with the stock cooler so the cooler purchase requirement point is an invalid one.
 
FrenchTart said:
You can hit a reasonable OC with the stock cooler so the cooler purchase requirement point is an invalid one.
Click to expand...
A reasonable overclock yes, but not high overclock to like 3.7-3.8GHz (which people need around 1.45-1.50 vcore) which is necessary to match the speed of i3 2100 in games.

And you are missing the point I was making. There's no reason people should build a new system based on AM3 with Phenom II X4 955/965BE NOW over a i3 2100, when:
a) build cost around the same for both
b) only deliver simliar frame rate when overclocked to 3.7GHz and in games that use four cores fully, but slower in games and general applications that don't
c) roughly twice the amount of power (TDP 140W at 3.7GHz vs TDP 65W)
d) hotter running (most people wouldn't even consider running Phenom II X4 at 3.7GHz or above with stock cooler)
e) the 1155 (P67/Z68) platform being more future proof and can upgrade to i5 2500K later, whereas the AM3 board is a deadend...and for the ones that's exception and are said to be able to "support" Bulldozer in the future, the overclock would most likely be rubbish on the old chipset board.

If you can provide a list of reasons why it would be better to get a Phenom II X4 build over a i3 2100 for someone who only use it for gaming and general usage, I would love to see it...

And regarding the Quad-core vs dual-core, I'll give you another example. When I play on Global Agenda, in areas there are not a lot of things my GPU would be at 99% usage and hitting 80-110fps, but when at demanding scenes with lots of things happening in open area, my frame rate would drop to as low as 23fps and with the GPU usage dropping from 99% down to 70%. This is clearly a CPU bottleneck...but why is it the case? Because the game only uses around 2-3cores. If I had a i3 2100, it would be able the same (higher in fact) power as all 4 cores of my Q6600 at 3.6GHz but over the two cores, so it will not be suffer at the hands of the lack of the game's support for the 4th core, and my GPU usage would most likely be able to hit 90-99% instead of the 70%, and getting higher frame rate as a result.

Extra cores is useless for games that doesn't make use of it, it is as simple as that.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom