Getting to the bottom of the bottleneck question.

[fffblackmage]

currently, my 4830 works pretty well with medium (sometimes high) detail settings and sometimes well enough with some AA. I will always play at 1920x1200, unless it's like that cryostasis demo... 1280x800 and still not even 30fps on average, but that was a weird game and poorly optimized. Source games run well 99% of the time, even though I'm using 8xAA with the max settings.

The first time I realized the CPU could be a bottleneck was when I was playing Painkiller on my old Athlon xp 2600+ with a GF4 TI4800. Overclocking my cpu from 1.9GHz to 2.3GHz made gameplay completely smooth. Mass Effect is also like that even with my E6320 + X1950XT. Unplayable at stock, but OCing the cpu made it playable.

 

[amenx]

Originally posted by: happy medium
Did you click the link at the top of my post and check out the benchmarks for 3d mark 2006? The guy went from a e8400 stock with dual 8800gtx's with 14,000 points to dual gtx 285's scoring 14,500.
If you look at his prior systems and scores he's been wasting money for years.

At this present time there are 4 post in video on the front page alone that deal with some kind of bottlneck question.

I believe it could help potential buyers of crossfire/ sli with weak cpu's also.

The G80's came out when 3dm06 was around, and Nvidia heavily optimized those cards to perform best with that benchmark. The G200 cards came out when Vantage had arrived, and they were optimized for THAT benchmark.

The G200 cards are KNOWN to perform poorly relative to the G80/92 cards in 3dm06.

NOW, look at vantage benchmarks comparing both G80/90 and the G200 cards and you will find a whopping difference between the two, even with dual core CPUs. The GTX285 in Vantage scores about double or more than a G80/92 card, even with the same CPU that you had erroneously assumed it was bottlenecked in 3dm06.

 

[cbn]

http://forums.anandtech.com/me...id=28&threadid=2304736

Speaking of CPU bottlenecks here is a thread I made today after looking at the Intel Roadmap. It looks like we are headed for eight core CPU as the default after "Sandy Bridge" (32nm processor)

 

[KevinH]

Originally posted by: SickBeast
IMO the CPU requirements are too steep in this guide. CPU limitations tend to be overblown.

I agree 100%.

It's always been the case...almost akin to the "power" requirements of systems.

Way overblown.

 

[nemesismk2]

Originally posted by: SickBeast
IMO the CPU requirements are too steep in this guide. CPU limitations tend to be overblown.

I agree 100% as well.

 

[sandorski]

There is too much "Bottleneck" talk, but this thread won't end it. :laugh:

 

[happy medium]

Originally posted by: SickBeast
IMO the CPU requirements are too steep in this guide. CPU limitations tend to be overblown.

Kenvin H, Nemesismk2.......

Cpu requirements too steep?

E8400 @ 3.0 vs i7 765 stock @ 3.2.....................

Brothers in Arms @ 1900x1200

4870 crossfire

62 fps vs 83 fps

4870 x4

63 fps vs 85 fps

260 sli....

55 fps vs 95 fps

280 sli ......

50 fps vs 97 fps

280 gtx 3x sli

53 fps vs 107 fps


http://www.guru3d.com/article/...e-performance-review/7

Call of duty 4 @ 1900x1200

e8400 @ 3.0 vs core i7 965 @ 3.2 @ 1900x1200

4870 crossfire

132 fps vs 124 fps

4870 x4

169 fps vs 219 fps

260 gtx sli

117 fps vs 117fps

280 sli

127 fps 132 fps

280 3x sli

145 fps 183 fps

http://www.guru3d.com/article/...e-performance-review/9

Far Cry 2 @1900x1200

4870 crossfire

50 fps vs 63 fps

4870x4

53 fps vs 94

260 sli

54 fps vs 77 fps

280 sli

55 fps vs 84 fps

tri 280's

52 fps vs 112 fps

http://www.guru3d.com/article/...-performance-review/11

Crysis Warhead @ 1900x1200

4870 crossfire

25 fps vs 32 fps

4870 x4

28 fps vs 35 fps

260 sli

38 fps vs 36 fps

280 sli

40 fps vs 50 fps

3x 280 gtx;s

35 fps vs 55 fps

http://www.guru3d.com/article/...-performance-review/15

3x 280 sli stock core i7 (3.2) vs overclocked @ 3.7

Brothers in arms

107 fps vs 115 fps

Call of duty 4

183 fps vs 183 fps

Far cry 2

112 fps vs 112 fps

F.E.A.R.

266 fps vs 266 fps

Crysis Warhead

55 fps vs 55 fps

http://www.guru3d.com/article/...-performance-review/15


E8400 @ 3.0 vs qx9770 @ 3.2 vs core i7 965 @ 3.2

Brothers in arms 1900x1200

53 fps, 67 fps, 107 fps

Far Cry 2 @ 1900x1200

52 fps, 77 fps, 112 fps

F.E.A.R.

229 fps , 250fps, 266fps

Crysis Warhead

32 fps , 42 fps ,55 fps

3d mark Vantage 280 tri sli

E8400 - 12219
QX9770 - 19057
i7 965 - 25070


No cpu bottleneck??????
Overblown? I think not.

 

[cbn]

Originally posted by: happy medium

Originally posted by: SickBeast
IMO the CPU requirements are too steep in this guide. CPU limitations tend to be overblown.

Kenvin H, Nemesismk2.......

Cpu requirements too steep?

E8400 @ 3.0 vs i7 765 stock @ 3.2.....................

Brothers in Arms @ 1900x1200

4870 crossfire

62 fps vs 83 fps

4870 x4

63 fps vs 85 fps

260 sli....

55 fps vs 95 fps

280 sli ......

50 fps vs 97 fps

280 gtx 3x sli

53 fps vs 107 fps


http://www.guru3d.com/article/...e-performance-review/7

Call of duty 4 @ 1900x1200

e8400 @ 3.0 vs core i7 965 @ 3.2 @ 1900x1200

4870 crossfire

132 fps vs 124 fps

4870 x4

169 fps vs 219 fps

260 gtx sli

117 fps vs 117fps

280 sli

127 fps 132 fps

280 3x sli

145 fps 183 fps

http://www.guru3d.com/article/...e-performance-review/9

Far Cry 2 @1900x1200

4870 crossfire

50 fps vs 63 fps

4870x4

53 fps vs 94

260 sli

54 fps vs 77 fps

280 sli

55 fps vs 84 fps

tri 280's

52 fps vs 112 fps

http://www.guru3d.com/article/...-performance-review/11

Crysis Warhead @ 1900x1200

4870 crossfire

25 fps vs 32 fps

4870 x4

28 fps vs 35 fps

260 sli

38 fps vs 36 fps

280 sli

40 fps vs 50 fps

3x 280 gtx;s

35 fps vs 55 fps

http://www.guru3d.com/article/...-performance-review/15

3x 280 sli stock core i7 (3.2) vs overclocked @ 3.7

Brothers in arms

107 fps vs 115 fps

Call of duty 4

183 fps vs 183 fps

Far cry 2

112 fps vs 112 fps

F.E.A.R.

266 fps vs 266 fps

Crysis Warhead

55 fps vs 55 fps

http://www.guru3d.com/article/...-performance-review/15


E8400 @ 3.0 vs qx9770 @ 3.2 vs core i7 965 @ 3.2

Brothers in arms 1900x1200

53 fps, 67 fps, 107 fps

Far Cry 2 @ 1900x1200

52 fps, 77 fps, 112 fps

F.E.A.R.

229 fps , 250fps, 266fps

Crysis Warhead

32 fps , 42 fps ,55 fps

3d mark Vantage 280 tri sli

E8400 - 12219
QX9770 - 19057
i7 965 - 25070


No cpu bottleneck??????
Overblown? I think not.

Thats interesting Crysis scales better with the Core i7 over the OC'd E8400 because I thought that game only used 2 CPU cores.

Or maybe it is because two cores of the i7 quad were not only faster in an absolute sense (3Ghz for the E840 vs 3.2 Ghz for the i7) but also clock for clock faster.

 

[cbn]

Originally posted by: happy medium

No cpu bottleneck??????
Overblown? I think not.

I think it depends on how many frame per second someone needs.

It seems most any dual core CPU will at least give 40 FPS if the resolution/detail settings are turned down (thus releasing any drag an insufficient video card would have on total computational time)

 

[happy medium]

What I was trying to show is with a e8400 @ 3.0 is not using all the gpu power of any high end sli/crossfire system @ 1900x1200.

Sometimes it's the difference between playable and unplayable.

If I invested in high end sli/crossfire system I would not want any game to dip below 35/40 fps. The most recent and upcomming games will be quad core optimized. Most duel core systems will only meet minimum system requirements.

I for one consider the recommended system requirements for a game to be the minimum requirements.

 

[apoppin]

Originally posted by: happy medium
What I was trying to show is with a e8400 @ 3.0 is not using all the gpu power of any high end sli/crossfire system @ 1900x1200.

Sometimes it's the difference between playable and unplayable.

If I invested in high end sli/crossfire system I would not want any game to dip below 35/40 fps. The most recent and upcomming games will be quad core optimized. Most duel core systems will only meet minimum system requirements.

I for one consider the recommended system requirements for a game to be the minimum requirements.

You need to realize that no one here wants to hear this. i have done plenty of this testing and it supports yours. Even for single powerful GPUs - starting with GTX280 - it is sometimes necessary to have a fast Quad for good playability at 19x12.

C2D is so last year

 

[heyheybooboo]

Originally posted by: happy medium
What I was trying to show is with a e8400 @ 3.0 is not using all the gpu power of any high end sli/crossfire system @ 1900x1200.

Sometimes it's the difference between playable and unplayable.

If I invested in high end sli/crossfire system I would not want any game to dip below 35/40 fps. The most recent and upcomming games will be quad core optimized. Most duel core systems will only meet minimum system requirements.

I for one consider the recommended system requirements for a game to be the minimum requirements.

This thread.


(no offense)

 

[AzN]

Originally posted by: apoppin

Originally posted by: happy medium
What I was trying to show is with a e8400 @ 3.0 is not using all the gpu power of any high end sli/crossfire system @ 1900x1200.

Sometimes it's the difference between playable and unplayable.

If I invested in high end sli/crossfire system I would not want any game to dip below 35/40 fps. The most recent and upcomming games will be quad core optimized. Most duel core systems will only meet minimum system requirements.

I for one consider the recommended system requirements for a game to be the minimum requirements.

You need to realize that no one here wants to hear this. i have done plenty of this testing and it supports yours. Even for single powerful GPUs - starting with GTX280 - it is sometimes necessary to have a fast Quad for good playability at 19x12.

C2D is so last year

That and there's only a handful of quad optimized PC games. C2D might be so 2006 but it still fast enough for the latest PC games.

 

[AzN]

Originally posted by: happy medium
Did you click the link at the top of my post and check out the benchmarks for 3d mark 2006? The guy went from a e8400 stock with dual 8800gtx's with 14,000 points to dual gtx 285's scoring 14,500.
If you look at his prior systems and scores he's been wasting money for years.

At this present time there are 4 post in video on the front page alone that deal with some kind of bottlneck question.

I believe it could help potential buyers of crossfire/ sli with weak cpu's also.

Who the hell plays 3dmark2k6?

 

[toyota]

Originally posted by: Azn

Originally posted by: apoppin

Originally posted by: happy medium
What I was trying to show is with a e8400 @ 3.0 is not using all the gpu power of any high end sli/crossfire system @ 1900x1200.

Sometimes it's the difference between playable and unplayable.

If I invested in high end sli/crossfire system I would not want any game to dip below 35/40 fps. The most recent and upcomming games will be quad core optimized. Most duel core systems will only meet minimum system requirements.

I for one consider the recommended system requirements for a game to be the minimum requirements.

You need to realize that no one here wants to hear this. i have done plenty of this testing and it supports yours. Even for single powerful GPUs - starting with GTX280 - it is sometimes necessary to have a fast Quad for good playability at 19x12.

C2D is so last year

That and there's only a handful of quad optimized PC games. C2D might be so 2006 but it still fast enough for the latest PC games.

yeah even clock for clock a quad core only provides a small advantage in a couple games and even then usually only at low res. at 1920 with any single gpu there isnt a playable advantage in any game that I know of except maybe GTA4.

 

[apoppin]

Originally posted by: Azn

Originally posted by: apoppin

Originally posted by: happy medium
What I was trying to show is with a e8400 @ 3.0 is not using all the gpu power of any high end sli/crossfire system @ 1900x1200.

Sometimes it's the difference between playable and unplayable.

If I invested in high end sli/crossfire system I would not want any game to dip below 35/40 fps. The most recent and upcomming games will be quad core optimized. Most duel core systems will only meet minimum system requirements.

I for one consider the recommended system requirements for a game to be the minimum requirements.

You need to realize that no one here wants to hear this. i have done plenty of this testing and it supports yours. Even for single powerful GPUs - starting with GTX280 - it is sometimes necessary to have a fast Quad for good playability at 19x12.

C2D is so last year

That and there's only a handful of quad optimized PC games. C2D might be so 2006 but it still fast enough for the latest PC games.

What if you play one of those games?

- i said SO 2007 .. not '06 when you were correct

yeah even clock for clock a quad core only provides a small advantage in a couple games and even then usually only at low res. at 1920 with any single gpu there isnt a playable advantage in any game that I know of except maybe GTA4.

you would be correct - last year

 

[toyota]

Originally posted by: apoppin

Originally posted by: Azn

Originally posted by: apoppin

Originally posted by: happy medium
What I was trying to show is with a e8400 @ 3.0 is not using all the gpu power of any high end sli/crossfire system @ 1900x1200.

Sometimes it's the difference between playable and unplayable.

If I invested in high end sli/crossfire system I would not want any game to dip below 35/40 fps. The most recent and upcomming games will be quad core optimized. Most duel core systems will only meet minimum system requirements.

I for one consider the recommended system requirements for a game to be the minimum requirements.

You need to realize that no one here wants to hear this. i have done plenty of this testing and it supports yours. Even for single powerful GPUs - starting with GTX280 - it is sometimes necessary to have a fast Quad for good playability at 19x12.

C2D is so last year

That and there's only a handful of quad optimized PC games. C2D might be so 2006 but it still fast enough for the latest PC games.

What if you play one of those games?

- i said SO 2007 .. not '06 when you were correct

yeah even clock for clock a quad core only provides a small advantage in a couple games and even then usually only at low res. at 1920 with any single gpu there isnt a playable advantage in any game that I know of except maybe GTA4.

you would be correct - last year

well what games show a playable difference at 1920 with say a gtx280? link me to some real numbers so I can see for myself because I am curious.

 

[apoppin]

i will send you a PM to a review - Crossfire with e8600 vs q9550s. You can follow the link to an earlier review with single GPU. Testing is at 19x12 and several games make a real difference with Quad over dual - never mind core speed

 

[toyota]

Originally posted by: apoppin
i will send you a PM to a review - Crossfire with e8600 vs q9550s. You can follow the link to an earlier review with single GPU. Testing is at 19x12 and several games make a real difference with Quad over dual - never mind core speed

okay thanks. nice work but your graphs are very confusing to look at. also I still came to the same conclusion. clock for clock at 1920 a quad still doesnt provide a real playable difference over a dual core of the same architecture. more times than not a faster dual core will beat a quad core unless I am having trouble reading graphs.

 

[apoppin]

Confusing, really? They are all color coded and you are the first one to mention it

it think it is pretty clear in few games where clock speed simply does not make up for "missing" 2 cores.
-look at FC2. Big differences! Quad kills dual across the board.

WiC is playable on a C2Q that boggs down completely on a C2D - at 19x12 resolution - even with a faster dual core!
ET-QW and FEAR all show Quad superior even at high resolutions with Crossfire-X3
So does UT3 and HL2

were you looking at the same review link that i sent you?

 

[toyota]

Originally posted by: apoppin
Confusing, really? They are all color coded and you are the first one to mention it

it think it is pretty clear in few games where clock speed simply does not make up for "missing" 2 cores.
-look at FC2. Big differences!
WiC is playable on a C2Q that boggs on a C2D - at 19x12 resolution
ET:QW and FEAR all show Quad superior at high resolutions with Crossfire
So does UT3 and HL2

were you looking at the same review i sent you?

yes I was looking at part 2 with the Q9550 and E8600 with the single card. your graphs are difficult because some of the colors are almost identical. also having dual cores with various speeds, quad cores with various speeds, and then two different gpus with all those combos hurts me eyes to read.

 

[apoppin]

yes - agreed; there is less difference with a single GPU - you are looking at Part 2, not Part 3
[the graphs are better in 3 also ]
- once you get to multi-GPU, the games that take advantage of it, really shine on a Quad over a dual

and this year we should have single GPUs as powerful as SLi or CF - that pretty much mandates Quad core
- the devs are clearly multi-threading for game performance imo - unlike 2 years ago

 

[happy medium]

Heres another good site for some answers to your bottlneck questions.

http://alienbabeltech.com/main/?p=4090

 

[toyota]

Originally posted by: happy medium
Heres another good site for some answers to your bottlneck questions.

http://alienbabeltech.com/main/?p=4090

lol

 

[AzN]

Originally posted by: apoppin
What if you play one of those games?

All 3 of them? Either I played it smoothly with my C2D or I didn't play it at all because the game sucked so bad.

http://alienbabeltech.com/main/?p=4090

Even in your own article you come up with minimal gains in majority of the games you tested. Again the games are not quad optimized other than slight better frame rates due to bigger cache on the q9550.