I experience half the FPS of one card with a crossfire R9 290 setup.

[Final_Fighter]

his performance should go up as new drivers are released, but i am a little curious with the process explorer to see if it will tell us anything more

i think power management has already been discussed but could double check but to me it looks like its working fine.

i dout that its another application because his cpu was idling fine when there was no load.

 

[OneMoar]

I am starting the wonder if he simply had a bunk connection between the cpu and socket and a reseat was all that was really needed
a 2500 is not a slow CHIP and we all know that a 2500 @stock will blow a phenom II out of the water so why is it that the numbers he was getting where WAYYYY below par I know that a phenom II and a pair of 5870's will push well over 70Fps in multi,1080p,high

 

[JaredzzC]

Well, I've no idea how to log the cycles in Process Explorer, so any help would be appreciated :/
I ran the intel processor diagnostic tool, and it passed the test (although I guess that doesn't tell us much):

I was going to run the Diagnostics test from 'Hot CPU Tester Pro 4', but it wants to do the test for 6 hours, so I'll leave that until later.

 

[RCoon]

I am starting the wonder if he simply had a bunk connection between the cpu and socket and a reseat was all that was really needed
a 2500 is not a slow CHIP and we all know that a 2500 @stock will blow a phenom II out of the water so why is it that the numbers he was getting where WAYYYY below par

Take a look at a couple of other forums. I said on page one, the OP was not the first person to have bottleneck issues with a 2500 + crossfire while playing crysis 3.

http://hardforum.com/showthread.php?t=1749286

http://forums.guru3d.com/showthread.php?t=381189

http://forums.steampowered.com/forums/showthread.php?t=2695876

Time to stop the witch-hunt eh?

 

[Final_Fighter]

just something i want to ask. can you check your bios and see if your ram is running at 1600mhz. i now when my bro put an i5 3570k in his board from a sandy i3 it bumped his ram speed up. he was using a z77 chipset like yours.

im just curious because that happened to my brother just want to now if its the same with you. should not effect performance in any negative way i think. might be wrong.

 

[newconroer]

I can now see that others are clearly pointing out what I've been saying from the start, so no need for me to reiterate that a WORKING i5 is not a massive bottleneck to two R9 290s.

I wish Jared had dual booted a fresh copy of his Windows and then tested it before getting the new CPU.

 

[JaredzzC]

just something i want to ask. can you check your bios and see if your ram is running at 1600mhz. i now when my bro put an i5 3570k in his board from a sandy i3 it bumped his ram speed up. he was using a z77 chipset like yours.

im just curious because that happened to my brother just want to now if its the same with you. should not effect performance in any negative way i think. might be wrong.

Still at 1333 no change

 

[Dent1]

Hi, Everyone.

I received my i7-3770k today and it runs like an absolute charm! (I haven't yet overclocked it)

To put it into perspective:
i5-2500 + 2x R9 290 was 30-50fps in Crysis 3 Multiplayer with all settings on the lowest.
The CPU usage was 100% with just 1GPU, so obviously it was 100% with 2GPUs.

i7-3770k + 2x R9 290 is 80-115fps in Crysis 3 Multiplayer with all settings on the highest(no AA).
The CPU usage peaks at around 90% with 2GPUs.

BF3's performance is very good, too. No where near as bad as what it used to be when using crossfire.

I am just as shocked as many of you are about the performance of my i5-2500. I thought it would be more than capable, but obviously not. It's hard to believe that the CPU would make that much of a difference in this case, but it did.
Before I bought my 2nd GPU I even asked around on multiple forums (Tom's Hardware etc) and I was told that the i5-2500 would be fine.

Maybe there was something wrong with the CPU? I've used it since March 2011, but still, its performance shouldn't have degraded.

I'd honestly like to thank everyone for their on-going help in trying to figure out the cause of my problem. It's greatly appreciated : )

What frame rate are you now getting with 1 GPU only? Is crossfire now outperforming it?

Then explain the OP's sudden miraculous double of FPS, if not to do with the change in CPU.

To be fair the OP hasn't yet specified whether his frame rate in crossfire is more than his frame rate in single card mode.

No point having your frame rate double in CF if it quadruples in single card mode. If this is the case the issue might not be fixed.

 

[Shambles1980]

i have no issues in believeing that a i7 can run a crossfire setup for those particular games better than an i5 can.
however i dont see why a i5 bottle neck would produce less fps in sli/xfire than it did with a single gpu.
the same fps i could understand. and only 2-3 more i could understand.
but to have half of the fps of a single card does seem to imply that there is more involved than just the cpu.
i wouldnt say that the cpu defiantly isnt the bottle neck. but something does seem a bit strange.

 

[Aquinus]

however i dont see why a i5 bottle neck would produce less fps in sli/xfire than it did with a single gpu.

That's because to do crossfire or SLI, there is some level of coordination overhead that is imposed on the CPU. So if you're already pegging out the CPU, adding another GPU will result in less performance. It shouldn't have been 50% less though, that's weird.

Either way, a new CPU seems to have solved the problem. Maybe we should just let it be? That sounds like a great plan to me.

Also with regard to the CPU contacts not making a good contact, I think that would cause problems beyond just slowing down GPU performance, such as BSODs or not booting at all.

 

[JaredzzC]

That's because to do crossfire or SLI, there is some level of coordination overhead that is imposed on the CPU. So if you're already pegging out the CPU, adding another GPU will result in less performance. It shouldn't have been 50% less though, that's weird.

Either way, a new CPU seems to have solved the problem. Maybe we should just let it be? That sounds like a great plan to me.

Also with regard to the CPU contacts not making a good contact, I think that would cause problems beyond just slowing down GPU performance, such as BSODs or not booting at all.

Exactly, that's what I thought. Since the CPU is locked into place so well, if the CPU wasn't seated, then surely either the pins on the CPU or the motherboard itself would be damaged and therefore hinder (quite badly) the performance of myCPU/PC wouldn't start up? When I placed my i7 in the socket, there didn't seem to be any damage whatsoever to the pins on either the CPU or the motherboard.

 

[JaredzzC]

I seem to have run into a "new" problem :/ I thought I had run the game both in Crossfire mode and single-gpu mode(hence why I posted the Crossfire results), but it seems I didn't.
I seem to be getting the exact same FPS with one card as I do with two cards...

Since I have swapped each GPU so that it has had time in the primary slot, each works fine, so I'm sure we can rule out a GPU issue ( but I'd be happy to run some tests, if you want ).
In terms of CPU, I ran the intel processor diagnostics test and it passed(not surprising since it's a new CPU).
I will be running the 'Hot CPU Tester Pro 4' to see if there are any other problems, memory etc (although it's very doubtful), but it's a 6 hour test so I probably won't reply for some time.
I can also run Prime 95, if you want.

Could it be a PCIe issue? E.g. the 2nd slot is not functioning properly? Is there anyway to test this, such as running 1 GPU from the second slot, rather than the first?

I apologise for misleading anyway as I genuinely thought I had tested both circumstances (single/multi-gpu mode), but at least my FPS is a massive improvement from before.

 

[fullinfusion]

290X != 290
also you have 3 cards the op has TWO and yes the scaling is horribly broken( with a single monitor) test it your self
your system is a completely different beast then the ops you have triple monitors and triple cards the op has a single monitor and 290's
don't believe me disable one card and unplug all but your primary monitor and run at 1080 or 1440p

Funny I play crysis 3 and I have no issues at all with scaling in crossfire on 2 290s... Butter smooth

CPU usage is WAY high for a 3770K you have another problem the new cpu is just masking it crysis 3 is not that CPU intensive
install proces explorer and logg the cycles used
http://technet.microsoft.com/en-us/sysinternals/bb896653.aspx

I agree and sorry to dp but on the phone
I had a 3770k and now 4790k and its hardly working at all at 4.5ghz and hardly even hits 40c on the hottest core.. And I run max ultra settings in all games.

 

[RCoon]

I seem to have run into a "new" problem :/ I thought I had run the game both in Crossfire mode and single-gpu mode(hence why I posted the Crossfire results), but it seems I didn't.
I seem to be getting the exact same FPS with one card as I do with two cards...

Since I have swapped each GPU so that it has had time in the primary slot, each works fine, so I'm sure we can rule out a GPU issue ( but I'd be happy to run some tests, if you want ).
In terms of CPU, I ran the intel processor diagnostics test and it passed(not surprising since it's a new CPU).
I will be running the 'Hot CPU Tester Pro 4' to see if there are any other problems, memory etc (although it's very doubtful), but it's a 6 hour test so I probably won't reply for some time.
I can also run Prime 95, if you want.

Could it be a PCIe issue? E.g. the 2nd slot is not functioning properly? Is there anyway to test this, such as running 1 GPU from the second slot, rather than the first?

I apologise for misleading anyway as I genuinely thought I had tested both circumstances (single/multi-gpu mode), but at least my FPS is a massive improvement from before.

During a synthetic benchmark, does afterburner show both GPU's loaded or just one now? If it's just one, that's a new issue entirely, as both were being loaded before correctly in benchmarks.

Also I assume you have the GPU's in the right slots for crossfiring?

You don't happen to have another PCIe card like a sound card or a USB 3.0 card plugged into one of your slots to do you?

 

[JaredzzC]

During a synthetic benchmark, does afterburner show both GPU's loaded or just one now? If it's just one, that's a new issue entirely, as both were being loaded before correctly in benchmarks.

Also I assume you have the GPU's in the right slots for crossfiring?

You don't happen to have another PCIe card like a sound card or a USB 3.0 card plugged into one of your slots to do you?

There are only 2 slots available for the GPU on the ASRock Z77 Extreme 4, and both are occupied by my GPUs

and there are no other PCIe cards inserted. no sound card/usb/wifi card

When I run the render test in GPU-Z, they seem to be near enough 100% usage with crossfire:

When I ran Furmark 15min Benchmark test, they were both around 100%, then it changed:

Weird, though. GPU 1 had a temperature of 94 degrees so it eventually throttled back (which made my score worse) and apparently it ran at 100%. GPU2 max temp was 83 and apparently the usage was all over the place:

Both GPUs seem to be fine, then. (At least according to afterburner). As far as I'm aware, possibly the CPU is fine (but I will still try and run tests since people think it's possible there is a memory leak), but that still doesn't explain the exact same FPS whether it's 1 card, or 2 :/
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Furthermore, in terms of 'memory leaks', nothing is screaming out at me here; it seems to be fine, but you may see differently:

 

[fullinfusion]

There are only 2 slots available for the GPU on the ASRock Z77 Extreme 4, and both are occupied by my GPUs

and there are no other PCIe cards inserted. no sound card/usb/wifi card

When I run the render test in GPU-Z, they seem to be near enough 100% usage with crossfire:

When I ran Furmark 15min Benchmark test, they were both around 100%, then it changed:

Weird, though. GPU 1 had a temperature of 94 degrees so it eventually throttled back (which made my score worse) and apparently it ran at 100%. GPU2 max temp was 83 and apparently the usage was all over the place:

Both GPUs seem to be fine, then. (At least according to afterburner). As far as I'm aware, possibly the CPU is fine (but I will still try and run tests since people think it's possible there is a memory leak), but that still doesn't explain the exact same FPS whether it's 1 card, or 2 :/
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Furthermore, in terms of 'memory leaks', nothing is screaming out at me here; it seems to be fine, but you may see differently:

Have you ran 3DMark and ran the Fire Strike test?
Run it in either Performance or Extreme, run one card then enable crossfire and run it again and tell me your scores.

 

[JaredzzC]

Have you ran 3DMark and ran the Fire Strike test?
Run it in either Performance or Extreme, run one card then enable crossfire and run it again and tell me your scores.

I ran the Firestrike via the Demo.
1 card: http://www.3dmark.com/3dm/3790152 (8889)
2 cards: http://www.3dmark.com/3dm/3790193 (13851)

 

[RCoon]

A GPU shouldn't throttle at 83 degrees due to heat, especially not when a GPU at 94 doesn't throttle! Crossfire obviously works in the first benchmark completely fine, and firestrike shows both GPU's work but you should probably be getting a higher score. I'm not entirely sure what to make of the furmark shenanigans. Out of curiosity, what are your temperatures on the motherboard north and sound Bridge?

 

[fullinfusion]

Yeah If I were you Id look here for a different Bios for your cards.

I'm scoring over 1500 points better then you with both cards. http://www.3dmark.com/fs/2591035

Don't worrie about my CPU OC as I was to lazy to down clock so just look at the gpu scoring

I bet ya a bios flash will help you out big time even at stock settings like me

 

[Dent1]

I seem to have run into a "new" problem :/ I thought I had run the game both in Crossfire mode and single-gpu mode(hence why I posted the Crossfire results), but it seems I didn't.
I seem to be getting the exact same FPS with one card as I do with two cards...

I called it first! This is why I was quick to dismiss RCoon when he said the issue is fixed. Aquinus also assume the issue was fixed. A frame rate increase in CF means nothing unless its increases beyond its single card performance.

The truth of the matter is your CPU was never bottlenecked.

I would create a new partition and install windows and Crysis 3 onto it and test it. This way you can eliminate the OS, drivers and user configuration.

 

[RCoon]

I called it first! This is why I was quick to dismiss RCoon when he said the issue is fixed. Aquinus also assume the issue was fixed. A frame rate increase in CF means nothing unless its increases beyond its single card performance.

Crossfire works in benchmarks. I'm done with this thread, too many people getting a little too angsty for my liking, and hardly anybody is reading the entire thread before posting.

Good luck to OP.

 

[Aquinus]

I called it first! This is why I was quick to dismiss RCoon when he said the issue is fixed. Aquinus also assume the issue was fixed. A frame rate increase in CF means nothing unless its increases beyond its single card performance.

He said it was working fine THEN saw that it really wasn't running in CFX. If you want to complain, complain to the OP misleading us. If he said it still wasn't working right (like it wasn't,) I wouldn't have said that. Clearly you haven't read the entire thread as @RCoon has eluded too. I too am about done helping if everyone is just going to get their panties in a bunch as well.

Hi, Everyone.

I received my i7-3770k today and it runs like an absolute charm! (I haven't yet overclocked it)

To put it into perspective:
i5-2500 + 2x R9 290 was 30-50fps in Crysis 3 Multiplayer with all settings on the lowest.
The CPU usage was 100% with just 1GPU, so obviously it was 100% with 2GPUs.

i7-3770k + 2x R9 290 is 80-115fps in Crysis 3 Multiplayer with all settings on the highest(no AA).
The CPU usage peaks at around 90% with 2GPUs.

BF3's performance is very good, too. No where near as bad as what it used to be when using crossfire.

I am just as shocked as many of you are about the performance of my i5-2500. I thought it would be more than capable, but obviously not. It's hard to believe that the CPU would make that much of a difference in this case, but it did.
Before I bought my 2nd GPU I even asked around on multiple forums (Tom's Hardware etc) and I was told that the i5-2500 would be fine.

Maybe there was something wrong with the CPU? I've used it since March 2011, but still, its performance shouldn't have degraded.

I'd honestly like to thank everyone for their on-going help in trying to figure out the cause of my problem. It's greatly appreciated : )

 

[Dent1]

He said it was working fine THEN saw that it really wasn't running in CFX. If you want to complain, complain to the OP misleading us. If he said it still wasn't working right (like it wasn't,) I wouldn't have said that. Clearly you haven't read the entire thread as @RCoon has eluded too. I too am about done helping if everyone is just going to get their panties in a bunch as well.

I agree the OP wasn't 100% clear which prevented a better response. But at the same time even in that quote he doesn't mention any testing for crossfire performance compared to single card performance which was a big clue not to assume it was fixed.

I was just pointing this out so other readers don't glance at your post or RCoon's post and prematurely think the issue has been resolved. It wasn't supposed to be a deliberate criticism.

 

[Aquinus]

I agree the OP wasn't 100% clear which prevented a better response. But at the same time even in that quote he doesn't mention any testing for crossfire performance compared to single card performance which was a big clue not to assume it was fixed.

I was just pointing this out so other readers don't glance at your post or RCoon's post and prematurely think the issue has been resolved. It wasn't supposed to be a deliberate criticism.

Touche, but he did say

i7-3770k + 2x R9 290 is 80-115fps in Crysis 3 Multiplayer with all settings on the highest(no AA).
The CPU usage peaks at around 90% with 2GPUs.

That was a vast improvement over what OP was getting before so the results sounded like it made sense. Going from a 2500 to a 3770k not overclocked shouldn't have made *that* much of a difference by itself with a single GPU which is why I assumed it was fixed. The numbers sounded right if the 2500 was getting bottlenecked on a single 290.

 

[Dent1]

Touche, but he did say

That was a vast improvement over what OP was getting before so the results sounded like it made sense. Going from a 2500 to a 3770k not overclocked shouldn't have made *that* much of a difference by itself with a single GPU which is why I assumed it was fixed. The numbers sounded right if the 2500 was getting bottlenecks on a single 290.

I agree with that part. He said his frame rate improved going from the 2500k to the 3770k. But because you said it was resolved I didn't want everyone to forget his single vs CF issue. I tried to press the OP for a response straight away with "What frame rate are you now getting with 1 GPU only? Is crossfire now outperforming it?". But it got mixed up in the shuffle. I was trying to avoid an early celebration.

But I apologise if it came across as rude it wasn't my intention.