weak quad vs strong dual

Page 4 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

2is

Diamond Member
Apr 8, 2012
4,281
131
106
If you're asking these questions you haven't read the thread. If you had, you'd also know I'm using a sandy bridge processor, specifically. Your variety if points are based on a dual core without HT (which we aren't talking about) vs a quad with turbo (which we aren't talking about) and games getting 3000-4000 fps (which we aren't talking about)
 

taltamir

Lifer
Mar 21, 2004
13,576
6
76
We are talking about those things, because they are reality unlike the OP's scenario which is fantasy.
And it doesn't say anywhere in the thread what games you are testing or WHY this comparison matters.
 

crashtech

Lifer
Jan 4, 2013
10,695
2,294
146
There's nothing wrong with the discussion of hypotheticals. I appreciate 2is's work in presenting real examples of the scenario laid out in the OP.
 

2is

Diamond Member
Apr 8, 2012
4,281
131
106
We are talking about those things, because they are reality unlike the OP's scenario which is fantasy.
And it doesn't say anywhere in the thread what games you are testing or WHY this comparison matters.

We aren't, you are. Seems to me you're substituting fantasy and reality depending on which one suits your needs. Like the 3000-4000 fps suggestion. Bottom line, the question was laid out in the OP. It was debated and now being tested. Thats the reality of this thread. You may not like it and you may think it's a waste of time, but it's my time to waste, and frankly I'm having fun with satisfying my and others curiosities.
 

TuxDave

Lifer
Oct 8, 2002
10,571
3
71
There's nothing wrong with the discussion of hypotheticals. I appreciate 2is's work in presenting real examples of the scenario laid out in the OP.

Yeah exactly this. It's refreshing to see a debate about theoretical scenarios evolve into actual testing. At least I find the results very interesting.
 

BrightCandle

Diamond Member
Mar 15, 2007
4,762
0
76
These tests were for the best case quad core weren't they? We need to see more average tests and dual core advantage ones as well for an even balanced test.
 

Maximilian

Lifer
Feb 8, 2004
12,604
15
81
Yeah exactly this. It's refreshing to see a debate about theoretical scenarios evolve into actual testing. At least I find the results very interesting.

This...

Im still all for team quad though :awe: But that's just me, i tend to play modern games more than older ones.

Its definitely not the "no brainer obviously the quad is better" situation I initially thought, dual with HT is putting up a good show!
 

2is

Diamond Member
Apr 8, 2012
4,281
131
106
I've updated post 59 with Far Cry 3 and Fallout 3 results.
 

taltamir

Lifer
Mar 21, 2004
13,576
6
76
There's nothing wrong with the discussion of hypotheticals. I appreciate 2is's work in presenting real examples of the scenario laid out in the OP.

The issue is that the OP's scenario is pure fiction

We aren't, you are. Seems to me you're substituting fantasy and reality depending on which one suits your needs. Like the 3000-4000 fps suggestion.

It isn't a suggestion, I have observed even higher FPS in games before and my only mention of them in this thread has been to say they are not relevant. But you are fixating on it for some reason.
 
Last edited:

2is

Diamond Member
Apr 8, 2012
4,281
131
106
These tests were for the best case quad core weren't they? We need to see more average tests and dual core advantage ones as well for an even balanced test.

I'm open to suggestions. What do you want to see?
 

crashtech

Lifer
Jan 4, 2013
10,695
2,294
146
The issue is that the OP's scenario is pure fiction
Not really, both 2is and I have been able to reproduce the conditions set in the OP. Perhaps it was fiction, but now we-1 are having some fun with the premise in real life.
 

taltamir

Lifer
Mar 21, 2004
13,576
6
76
Not really, both 2is and I have been able to reproduce the conditions set in the OP.
Being able to artificially create it doesn't mean it isn't a fictional scenario.

I can create a test bed for a 0.5ghz single core i7 ivy bridge with HT enabled by disabling cores and manipulating the frequency of other products (heck, I actually have done it in the past).

The scenario is fictional because no such product exists on the market for me to buy, only emulate by manipulating actual products for some "just for fun" testing

If all you are doing is faffing about just for fun then I misunderstood and I apologize for being such a wet blanket
 
Last edited:

2is

Diamond Member
Apr 8, 2012
4,281
131
106
We get it, you think it's pointless and a waste of time. We don't care. You've provided your 2 cents, now if you'd kindly stop thread crapping, that would be greatly appreciated.
 

crashtech

Lifer
Jan 4, 2013
10,695
2,294
146
Being able to artificially create it doesn't mean it isn't a fictional scenario.
I suggest you first stop posting in this thread, since it doesn't meet your standards, and second, review the meaning of fictional. Pointless might be the word you were searching for, I don't know, but regardless, you are making a subjective judgement.
 

taltamir

Lifer
Mar 21, 2004
13,576
6
76
We get it, you think it's pointless and a waste of time. We don't care. You've provided your 2 cents, now if you'd kindly stop thread crapping, that would be greatly appreciated.

No, you do NOT get it and don't go assuming what I think. I misunderstood the situation (I thought the OP wanted to compare actual real products and made a mistake on what is available on the market; not a just for fun comparison of non existing hypothetical products) and have gone above and beyond mere civility by actually apologizing for that misunderstanding. You need to take a couple of chill pills and stop being so hostile and aggressive

I would have understood faster had you actually explained instead of attacking all the time
 
Last edited:

2is

Diamond Member
Apr 8, 2012
4,281
131
106
Apology accepted

Back to our regularly scheduled program... I'm looking forward to testing Bordelands 2 with an AMD card and the different level of Physx (low, med, high). Physx can still be enabled in this game with an AMD card except the additional effects are rendered on the CPU instead of the GPU, and from what I've read, only on a sinlge thread
 

cytg111

Lifer
Mar 17, 2008
26,216
15,624
136
Maybe it is not that easy to deduce.

All other things being equal and minimal kernel times(context switching and cache trashing) we could argue that a dual core 2GHz equals a 1Ghz quad core for binaries that is tasking 4 threads equally hard.

However realworld scenarios often does not fit that bill, with the dualcore, in theory, you would have the total output of that of the quad core with the option to run a singlethread at double speed.
That seems like a win.
But what about scheduling? Suppose you have 4 threads and one of them takes realtime priority(on the dual), you would have one thread bound to one core 'overperforming' and 3 threads on the second core lagging behind.. that wouldnt be an issue on the quad.
To me it seems like a match between the developers targeting n cores and a healthy relationship with the schedular on the target platform.
OR its a brainfart! Pffff....
 

2is

Diamond Member
Apr 8, 2012
4,281
131
106
I don't think there's much, if any doubt that true multi-tasking, particularly when there's a thread hogging up an entire core, the quad would be the clear winner.

One thing I've come to appreciate more after running these tests is Intel's decision to target efficiency over pure performance gains. Even when castrating this 2600k to half it's stock potential, every game I've tested so far was easily playable. BF3 could have used more performance, but it was hardly unplayable. The 14 fps minimum recorder for Fallout 3 I think was due to a save/load point than an actual drop in the game, because I never noticed it.
 

crashtech

Lifer
Jan 4, 2013
10,695
2,294
146
Since I've owned an Ivy i3, I've become pretty impressed with its efficient use of resources.
 

BrightCandle

Diamond Member
Mar 15, 2007
4,762
0
76
Games I know don't show much in the way of scaling between 2 cores and 4 cores (from gamegpu.ru that you haven't already tested):

Wargame Airland Battle
Call of Juarez Gunslinger
World of tanks 0.8.5.0
Sim City
World of warplanes
Starcraft 2
Defiance
Dead space 3
Shootmania Storm
Warframe
Mechwarrior online
Hawken
Assassins Creed III
Hitman Absolution
Call of Duty Black Ops II
Project cars
Of orcs and men
Dishonoured
Xcom enemy unknown
Black mesa
Guild wars 2

Some number of those will likely show cases where the dual core will exceed the quad or even it.
 

Maximilian

Lifer
Feb 8, 2004
12,604
15
81
Games I know don't show much in the way of scaling between 2 cores and 4 cores (from gamegpu.ru that you haven't already tested):

Wargame Airland Battle
Call of Juarez Gunslinger
World of tanks 0.8.5.0
Sim City
World of warplanes
Starcraft 2
Defiance
Dead space 3
Shootmania Storm
Warframe
Mechwarrior online
Hawken
Assassins Creed III
Hitman Absolution
Call of Duty Black Ops II
Project cars
Of orcs and men
Dishonoured
Xcom enemy unknown
Black mesa
Guild wars 2

Some number of those will likely show cases where the dual core will exceed the quad or even it.

With some of the custom 10v10 servers on a tiny map where everyone starts with a shit ton of points (and by extension units) those with a stock Q6600 or less get booted out for lagging :D It can make great use of a quad under the right circumstances.
 

Enigmoid

Platinum Member
Sep 27, 2012
2,907
31
91
Games I know don't show much in the way of scaling between 2 cores and 4 cores (from gamegpu.ru that you haven't already tested):

Wargame Airland Battle
Call of Juarez Gunslinger
World of tanks 0.8.5.0
Sim City
World of warplanes
Starcraft 2
Defiance
Dead space 3
Shootmania Storm
Warframe
Mechwarrior online
Hawken
Assassins Creed III
Hitman Absolution
Call of Duty Black Ops II
Project cars
Of orcs and men
Dishonoured
Xcom enemy unknown
Black mesa
Guild wars 2

Some number of those will likely show cases where the dual core will exceed the quad or even it.

GW2 shows minor gains from 2 to 4 cores. Bascially Gw2 has three main threads and will scale as such.

Hitman also tends to be CPU heavy.

hitman%20proz%202.png


Looks like it uses 6 cores.
 
Last edited:

2is

Diamond Member
Apr 8, 2012
4,281
131
106
Games I know don't show much in the way of scaling between 2 cores and 4 cores (from gamegpu.ru that you haven't already tested):

Wargame Airland Battle
Call of Juarez Gunslinger
World of tanks 0.8.5.0
Sim City
World of warplanes
Starcraft 2
Defiance
Dead space 3
Shootmania Storm
Warframe
Mechwarrior online
Hawken
Assassins Creed III
Hitman Absolution
Call of Duty Black Ops II
Project cars
Of orcs and men
Dishonoured
Xcom enemy unknown
Black mesa
Guild wars 2

Some number of those will likely show cases where the dual core will exceed the quad or even it.

I did test Black Ops II. Out of the other games, AC3 and Dead Space 3 are the only other ones I have. I have Sim City 3 too but I've never played it and assume I'd have to get a decent amount of stuff going before there would be much load.