EK Supreme HF Orientation & Performance

[zalbard]

Yes, I know that horizontal orientation (flow from left to right) is the best one for performance.
However, I am considering something different due to tube routing.
I can swear I've seen a thread with someone testing EK Supreme HF in all 4 orientations and posting the results. I can not find it, though, and I spent the last 2 hours looking for it.
Mostly interested in the performance of the block rotated by 90 degrees counter-clockwise (flow from bottom to top, should be pretty restrictive, and I recall like 0.5C performance hit), and 180 degrees (although I assume the latter should have no performance hit).
Feel free to help me find that thread (or review), or post your experiences. Thanks!

[tiborrr]

Sandy Bridge die has a obvious rectangular- and not square shape as the Ci7 predecessor. Maybe you would see a tiny gain under load if you would mount the waterblock rotated 90° (so the EK branding is rotated 90°). It's hard to say - if the block is mounted in a normal position the die would be covered with more channel. If the block is rotated 90° then the length of channels passing over the die is longer. All in all there shouldn't be a measurable difference.



Like this:


Regards,
Niko

[penguins]

as long as the flow is correct the position the block is in doesn't really matter if i recall correctly.

[Church]

tiborrr: do you happen to know, how lga2011 hexa core cpu die is supposed to look and what orientation might be preferred for them? Similar to SB?

[penguins]

Good point Churchy, not all of us have upgraded : )

heck for that matter.. what about AMD's?

[zalbard]

Thanks for your replies!

Like this:

Yep, that's the one I'm talking about (90 degrees counter-clockwise).

Sandy Bridge die has a obvious rectangular- and not square shape as the Ci7 predecessor. Maybe you would see a tiny gain under load if you would mount the waterblock rotated 90° (so the EK branding is rotated 90°). It's hard to say - if the block is mounted in a normal position the die would be covered with more channel. If the block is rotated 90° then the length of channels passing over the die is longer. All in all there shouldn't be a measurable difference.

I actually don't have SB, was hoping to sort out my i7 920 rig.
Either way, I am correct thinking that there should be little to no difference, right?

[tiborrr]

Correct. Just rotate the waterblock in the direction which allows you the easiest tubing routing!

@churchy: Don't know yet, haven't seen the die photos yet.

[PiLsY]

I actually managed to up and reversed the flow direction as well as rotating the mount (mount change was intentional to test a new tubing run). It was only 2c hotter with the flow direction reversed through the block and the 90 degree rotation.

As Tiborr guessed the difference is negligible.

[Russ_64]

When you guys say, Left to Right does that not depend on the way the chip/socket sits? What I mean is that in my DFI the chip bottom is at the top (of the socket) - so should I turn the block to run Right to Left?


At present I have my LT block running Top to Bottom but I am replacing it with a Full Copper HF this week - does the new "Easy Mount" mean I don't need the EK Back-plate (s1366) or can I still choose to use the "old" mounting? Just hope it has plate #1 installed as I do not fancy opening a new block...........