What's normal temp for GTX590?

[Kung Pow]

Well a 240mm rad approximately has the ability to dissipate 450Watts.
A stock GTX 590 at full load would get you 330Watts according to reviews, when overclocked it could reach up to 400 Watts.

So the radiator definitely is not the problem, at least from a performance point of view.
This loop is fairly small, so even if the D5 would impose a problem I do not think it would be noticable in such a small loop, so letīs cancel that too, but to be safe check if the flow is reasonable so you can positively strike out the pump.

My guess is the GPU block.
Did you assemble it yourself?
I would disassemble the GPU block, check it, pads etc and reassemble.

Also make sure you have the latest drivers installed, and check if those insane temps also appear with other games, benchmarks and stess tests.

[jv916]

The pump is more than sufficient for the application. I can see the fluids pushed through the tubes rather quickly. I assembled the block myself. I made sure all thermal pads are in the right spots. I've already dismount the block and remount once already. All other games (Crysis 2, COD Modern Warfare 3, Skyrim Elder Scroll V, NFS Shift 2) have similar results.

[Deton]

I know it is due to pump/rad location but it is better to have the inlet coming from the left port instead from the right (as you have it right now). This is mean you get a lot of heat dump from VRMs first before cooled coolant reaches the GPUs.

[Kung Pow]

What are the GPUīs Idle temps?
What is your ambient temperature?
What is the water temperature?
Fan RPM?
Have you bled the system properly? (this certainly could not be the main factor, but it contributes to bad temps)
What were your temps with the card on the reference cooler?

If the thermal paste was applied properly to the chip, and the rest of the block has also been mounted correctly, I donīt see any other option but to RMA the block, since all other possibilities have already been positively negated of being responsible.
Are you sure it has optimal contact to the card?

[Deton]

I know it is due to pump/rad location but it is better to have the inlet coming from the left port instead from the right (as you have it right now). This is mean you get a lot of heat dump from VRMs first before cooled coolant reaches the GPUs.

Kung Pow; he's already posted the temp, rpm...etc, few post back.

[jv916]

What are the GPUīs Idle temps?
What is your ambient temperature?
What is the water temperature?
Fan RPM?
Have you bled the system properly? (this certainly could not be the main factor, but it contributes to bad temps)
What were your temps with the card on the reference cooler?

If the thermal paste was applied properly to the chip, and the rest of the block has also been mounted correctly, I donīt see any other option but to RMA the block, since all other possibilities have already been positively negated of being responsible.
Are you sure it has optimal contact to the card?

GPU1 idle temp is 37
GPU2 idle temp is 33
ambient temp is 22
water temp is 32
fan rpm is +/- 1500 rpm
pump is running at 4.5 speed

I bled the system properly, tilted the case is all possible ways, release all the air bubbles from the reservoir and add more fluid until there is small amount of bubbles left in the system. When I dismounted the block, I did notice that the block did not make full contact in all surface areas of the GPUs. It seems as though there were more surface contact from slight center to one side of the GPU surface. The other half side did not even touch the block. When I remounted the block I made sure to put in the 4 screws around each GPU and tighten them down first then put in the rest of the screws. I don't know if the block is flawed or they are all the same.

[Kung Pow]

I only saw load and idle:P

The thing with the left inlet is common, a lot of people think that it matters.
I personally did not experience any differences with my GTX 295 blocks.
The water still utilizes and ultimately the premature VRM heat dump could only yield marginal temp losses regarding the core temp of the chip.

It is almost the same thing as saying that pump->rad->block is better than pump->block->rad ...if measruable, the gains are marginal, arenīt they?

[jv916]

I know it is due to pump/rad location but it is better to have the inlet coming from the left port instead from the right (as you have it right now). This is mean you get a lot of heat dump from VRMs first before cooled coolant reaches the GPUs.

Kung Pow; he's already posted the temp, rpm...etc, few post back.

Right now the flow goes from reservoir/pump, radiator, GPU block, then back to resevoir/pump. I believe this is the correct flow, I may be wrong. But how is changing the inlet to the left port going to change temperature. Does it really matter which port is inlet and outlet on this radiator? Is there a flow direction design to this radiator?

[Kung Pow]

GPU1 idle temp is 37
GPU2 idle temp is 33
ambient temp is 22
water temp is 32
fan rpm is +/- 1500 rpm
pump is running at 4.5 speed

I bled the system properly, tilted the case is all possible ways, release all the air bubbles from the reservoir and add more fluid until there is small amount of bubbles left in the system. When I dismounted the block, I did notice that the block did not make full contact in all surface areas of the GPUs. It seems as though there were more surface contact from slight center to one side of the GPU surface. The other half side did not even touch the block. When I remounted the block I made sure to put in the 4 screws around each GPU and tighten them down first then put in the rest of the screws. I don't know if the block is flawed or they are all the same.

The watertemp at least shows that the block is transferring a certain amount of heat to the water, so some contact had to be there.

Make sure the block is properly and tightly making full contact to the card though!

What is the watertemp under load?

[Irocing]

As to the inlet or outlet on the Rad, makes no difference as to flow or restriction.

Later

[jv916]

Water temp is 47 at full load. Somehow I have a gut feeling the radiator is not sufficient. The radiator feels pretty hot to the touch under full load.

[Kung Pow]

Water temp is 47 at full load.

In accordance to your load core temps, your water temp does not add up.
Typically that is the case, when blocks do not transfer enough heat off the components, due to incomplete or faulty mounting or just impeded contact.
And it couldnīt be just dissipated by the rad, because you only have 50Watts of additional open space dissipation ability with your 240 in regard to the 590īs heat dump.

Since you checked the block, twice and everything else seems to be alright, I would say RMA time, unless anybody else has something to say to this still?

You havenīt told me yet what your temps were like on the reference air cooler of your 590, those values might help as determine if it is really the block or actually a problem with the graphics chip itself.

[jv916]

In accordance to your load core temps, your water temp does not add up.
Typically that is the case, when blocks do not transfer enough heat off the components, due to incomplete or faulty mounting or just impeded contact.

Since you checked the block, twice and everything else seems to be alright, I would say RMA time, unless anybody else has something to say to this still?

You havenīt told me yet what your temps were like on the reference air cooler of your 590, those values might help as determine if it is really the block or actually a problem with the graphics chip itself.

With the stock air cooler, GPU1 idles at 39-40, GPU2 idles at 37-38. Full load GPU1 is at 78-80 and GPU2 is at 75-78. It seems like it didn't make much difference going to water cooling.

[Kung Pow]

With the stock air cooler, GPU1 idles at 39-40, GPU2 idles at 37-38. Full load GPU1 is at 78-80 and GPU2 is at 75-78. It seems like it didn't make much difference going to water cooling.

I think that concludes it.

Obviously something is wrong with your block.
Up until this point we have established evidence for every other component in your loop to be operating properly, except for the block.

The watertemp, previously mentioned contact issues, etc etc all indicate a problem with your block.
Disassemble the block, and RMA it.

And just fyi, 78 degrees at load is actually phenomenal for a 590 with stock heatsink lol

[jv916]

It seems the block is doing a great job by transferring heat to the coolant. To me, the radiator is just not sufficient and doesn't cool down the fluid fast enough. After 5-10 minutes at full load and touching the block, tubing, fittings, and reservoir it feels pretty hot. You don't think adding a bigger radiator to the loop will bring down the temp? The last thing I wan to do is RMA the block if that isn't really the problem.

[Kung Pow]

It seems the block is doing a great job by transferring heat to the coolant. To me, the radiator is just not sufficient and doesn't cool down the fluid fast enough. After 5-10 minutes at full load and touching the block, tubing, fittings, and reservoir it feels pretty hot. You don't think adding a bigger radiator to the loop will bring down the temp? The last thing I wan to do is RMA the block if that isn't really the problem.

It is my guess, based on the most logical solution and explanation regarding your loop.

A common 240 radiator is capable of dissipating 450 Watts.
The Black Ice is even a little more performant.

I have already shown to you how much Wattage comes off a 590 both at stock and overclocked based on reviews, and your radiatorīs dissipation capability is even more than sufficient by an amount of 50Watts.

Of course the radiator could be mailfunctioning, due to internal failure, which is unprobable to be honest, but strictly considering your radiatorīs specs and dissipation capability in theory, there should not be a problem.

Before you stated, that the block was not making full contact.
Also considering your load Watertemp in comparison to the core temp of the chip and the radiatorīs dissipation influence, the values indicate a rather improper heat transfer from the block to the water.

All I can do is give you a hint based on personal experience, you will have to make the decision.
If I were in your place I would have the block checked, because it is the only part indicating, tendencies to failure in this case, at least according to your statements.