Sign in with
Sign up | Sign in
Your question

Gaming performance wise, which computer is better?

Last response: in Video Games
Share
November 30, 2012 12:50:53 PM

So yeah, this generic question again... sorry if I'm in the wrong thread.

Both computers I'm looking at. I'm looking around for a decent computer under a thousand, and these caught my eye. I know, broad spectrum - laptop and desktop, but these are the ones I'm looking at; both around 1 grand.


3rd Generation Intel® Core™ i5-3450 Processor (3.10 GHz, 6MB Cache, with Turbo Boost Technology 2.0)
Windows® 7 Home Premium SP1 64 bit (English)
6GB 1600MHz DDR3
1TB SATA 3.0Gb/s Hard Drive with Native Command Queuing
1GB GDDR5 NVIDIA® GeForce® GT 545





3rd gen Ivy Bridge i7
Nvidia GTX 660M
(I don't know the full specs, sorry)

So basically what I'm concerned about is gaming performance and speed. Right now I want to leave hard drives and ram out of the equation since I'm on a budget.

Here's what I figured out so far:

i7 beats i5 fairly easily, no fuss.

The GTX660M seems to have a greater advantage over the GT545... it seems? That's what I thought; higher number, better performance? Unless the M for mobile taxes it greatly.

So which is better? I know there is no clear answer, but from the specs which looks like in theory it'd perform better?

Thankyou so much for your time.
November 30, 2012 12:57:48 PM

First one is better. And laptop i7 doesn't beat desktop i5. By a long shot. Desktop one is three times faster (and I'm not exaggerating).
m
0
l
November 30, 2012 1:21:21 PM

Thanks for the info, but could you please explain why and how? I don't need oversimplified layman's terms: I just want to know why.
m
0
l
Related resources
November 30, 2012 1:27:28 PM

laptop cpu's are hobbled and slower than desktop at a model number vs model number level. 2.2 ghz mobile ivy = 2.2 ghz desktop ivy. Mobiles have a lower upper limit and desktops start above that limit.
m
0
l
November 30, 2012 1:31:19 PM

Laptops are highly over inflated price wise as they fit more into a smaller space but they tend to cost 1.5-2x more for the same performance and they can't perform the same as high end desktops.

Both of the rigs, however, are very poor value for money.
m
0
l
November 30, 2012 2:17:57 PM

icalebkim said:
Thanks for the info, but could you please explain why and how? I don't need oversimplified layman's terms: I just want to know why.


Laptops cannot get same cooling and electricity power that a desktop will. That's why the manufacturers of CPUs have to reduce its power consumption. There's only one way to do that: alter architecture a bit, and in result, they become much slower. Power saving hurts. At the same frequency, desktop CPUs are around 1.5-2.0 the speed of laptop ones, and desktop CPUs run at higher frequency.

As for i5 or i7 - it's just a label, it doesn't really mean anything.
m
0
l
November 30, 2012 3:33:25 PM

hafijur said:
The laptop cpu is around 30-50% faster then the desktop cpu in highly threaded tasks as it is a i7 3630qm but generally around 10-20%. To be honest the desktop i5 has literally no chance to get close to the i7 unless you oc the heck out of the desktop cpu. People like you make me laugh. Also the laptop 660m is around 2x more powerful then a 545.


3630qm, stock speed = 2.4Ghz, turbo = 3.4 Ghz but limited by thermal solution, may never reach it, 3450 = 3.1 stock turboing to 3.5 with a thermal solution that can cope will be much much faster. so your first point is dubious.

your second point...
there is about a 25% increase from 545 to 660M http://www.videocardbenchmark.net/video_lookup.php?gpu=... I don't like passmark but they do at least compare mobile and desktop. Note again that the thermal solution for the 660m will have an impact, note that a 660m is not the same as a 660.
http://www.geforce.com/hardware/notebook-gpus/geforce-g... 1/3 as many cores as...
http://www.geforce.com/hardware/desktop-gpus/geforce-gt...
http://www.geforce.com/hardware/desktop-gpus/geforce-gt... no great but not half the speed.

you really make me laugh

m
0
l
November 30, 2012 3:37:34 PM

Passmark is like the WORST tool to benchmark anything. It thinks GTX 680 is better than GTX 690. So yea, don't trust that benchmark. How can i7-3630QM, which eats half power and has only 60% of the clock speed compare to desktop i5? I've tested them myself: the laptop i7 (i7-2670QM) was 3.64 times slower than my desktops i5-2500k. And where is your info that 660m is more powerful than 545? It has very similar performance according to notebookcheck.net

P. S. by that laughing comment you made yourself look like a fool.
m
0
l
November 30, 2012 4:10:15 PM

Sunius said:
Passmark is like the WORST tool to benchmark anything. It thinks GTX 680 is better than GTX 690. So yea, don't trust that benchmark. How can i7-3630QM, which eats half power and has only 60% of the clock speed compare to desktop i5? I've tested them myself: the laptop i7 (i7-2670QM) was 3.64 times slower than my desktops i5-2500k. And where is your info that 660m is more powerful than 545? It has very similar performance according to notebookcheck.net

P. S. by that laughing comment you made yourself look like a fool.


dude agreeing with you, declaring hafijur as being wrong. i'd say 545 and 660m are similar.
m
0
l
November 30, 2012 4:12:05 PM

By the way just to inform about turbo: it works on single threaded applications only. Once more than 1 core gets load on it, turbo goes away. So hafi, I hope you won't use that argument..
m
0
l
November 30, 2012 5:03:18 PM

in a laptop the thermal element of turbo 2.0 will clamp the speed down, especially as this is most likely to happen when the gpu is loaded hence overloading the thermal solution, probably capping it at stock.
m
0
l
November 30, 2012 6:11:41 PM

Exactly what 13thmonkey said. That's what happens in practice. Turbo is just a market phrase.

And no way i7-3630QM will win. 8 threads mean nothing. It doesn't give extra performance at all: all it does is if there are more than 4 threads active, it divides each cpu core to 2 parts and lets it do one thread each. I've tested that myself: the CPU benchmark I used lets you set the amount of threads to test. 4 thread test had exact same result as 8 thread test. This was done on i7-2670QM.
m
0
l
November 30, 2012 6:32:30 PM

I told you not to look at PassMark. It's inaccurate at best. And logic/way technology works is on my side.

As I said, I've benchmarked one myself. If there is somebody with i7 and wants to contribute - we could make another benchmark.
m
0
l
November 30, 2012 7:42:07 PM

At this moment I do not own a laptop with an i7. Though I used Intel Burn Test for benchmarking, standard stress level, 5 times to run. I get 102 GFlops on my i5-2500k today, and at a time of benchmarking the i7-2670QM, it got 28 GFlops.

You can download it from here:

http://www.softpedia.com/get/System/Benchmarks/IntelBur...
m
0
l
December 1, 2012 7:47:37 AM

My 2500k is at 4.3 GHz atm. At stock settings I get 85 GFlops. I really don't know what to say really: that's the scores I got when benchmarking myself. I might have access to that laptop today and I'll recheck it.

m
0
l
December 1, 2012 5:35:56 PM

Well, doesn't this mean that just that i5-2500k is superior? i5-3450 has AVX too. So does the laptop ones.

Anyway, had a chance to retest the laptop with i7-2670QM today. Cleaned it from dust and reapplied thermal paste (it was very dusty). To my surprise, speeds increased: got 48-50 GFlops. Will post a screenshot later.
m
0
l
December 2, 2012 12:09:43 AM

First one Is way better!

The second one is kinda good, but can't handle all the games.

The first one can probaly handle bf3 better than my pc.

I can only run it 10-20 fps fmll ;-;
m
0
l
December 2, 2012 9:13:09 AM

hafijur said:
Well, doesn't this mean that just that i5-2500k is superior? i5-3450 has AVX too. So does the laptop ones.
.........................................
Considering an i5 2320 scores 44.1 gigaflops, I doubt a 2500k is much faster. Do a non avx test on your 2500k at stock if you can.


Any idea how to disable it? Also, i7-2670QM scored 48-50 GFlops with AVX enabled. The scores were identical with 4 and 8 threads.
m
0
l
December 2, 2012 10:23:24 AM

and it can't be upgraded at all as it is a laptop that will be severely thermally constrained, and will be throttled often.
m
0
l
December 2, 2012 12:28:22 PM

Not really, we still didn't conclude our argument about the CPU and GPU.

Firstly, if my i5 gets 85 GFlops on stock settings, so will i5-3450, even if it's 200 MHz lower on clock - the ivy bridge architecture guarantees that. If i7-2670QM gets 50 GFlops, there's absolutely no way i7-3630QM is 70% faster...

Secondly, there is absolutely no proof that GTX 660M is 20% faster than GT 545. According to notebookcheck.net, GTX 660M is between GTS 450 and GTX 550 Ti in performance. So is GT 545, so their performance is very close.

And lastly, you cannot assume you can upgrade a laptop. On the contrary, one should assume you cannot. And that is why desktops are superior.
m
0
l
December 2, 2012 1:42:44 PM

I'm not gonna pay for benchmark to prove my point. Geekbench isn't free.

I still see no option to turn off AVX. Anyway, why would one want to do that? You're gaming and all with it enabled, and since all the desktop AND laptop CPUs have it, isn't just comparing it to each other with AVX on be fair? The benchmark I've done on i7-2670QM with 50 GFlop result was with AVX on. On stock, I get 85 GFlops on my i5. I still see no way how i7-3630QM could surpass it.

And lastly, there are countless benchmarks that show that i5 isn't any worse than i7 for gaming. So why are we arguing again? It's apparent that the desktop i5 will have higher clock speed, therefore will be faster for games.
m
0
l
December 2, 2012 1:47:03 PM

Try Performance Test Benchmarking it's a program that benchmarks your pc, and it's free ( trial 30 days )

All the best ~ Darkky.
m
0
l
December 2, 2012 2:56:49 PM

But the problem is they aren't at same clock speed :) .
m
0
l
!