Quote:
Originally posted by Blackeagle
The temps I posted were based on info posted by Cathar. He had 2c better with his MD30rlzt Iwaki pump. That is a really xtremly high pressure pump.
|
With the Eheim 1048 & 1250 as well, not just the MD-30RZ. Improvement ranged between 1.5°C and 2.0°C
Both blocks had been lapped using the same procedure. Heat load was a T'Bred B CPU running at 2400MHz/1.85v, running BurnK7 as a heat stress program.
Results were verified later to apply to the Barton CPU's as well.
P4's are just weird with their thermal probe which seems to be more indicative of the IHS temp, rather than the die temp. The P4 thermal probe doesn't actually monitor the CPU die temperature for purposes of determining the peak CPU temperature for thermal throttling and shutdown, so anything that P4's report is just random pot-luck. I did see 1.5°C lower temperatures on my crappy 2.0A Northwood running at 2.5GHz/1.65v.
Overclocking the Barton CPU very highly (2.70GHz/2.10v which was stable for all blocks tested when running BurnK7) the performance gap between the White Water and the Cascade was well over 2.5°C (not quite 3°C), however I've always been hesitant to report that since people will tend to run away and say that the Cascade is 3°C better than the White Water, then stick one onto their moderately overclocked system, and then wonder why they see much smaller differences.
The Cascade SS pulls between 0.5-1.0C clear of the copper Cascade for moderate overclocks, and a full 1.0C ahead for very high overclocks. The silver block was always pulling higher stable overclocks (20-40MHz depending on CPU used) than the copper block, which I feel is trying to tell a greater story than what simple temperatures alone can say.
Let us state again that these results are just what I see. Your mileage may vary depending on a great number of factors.
I believe that Phaestus is almost done with his copper Cascade review, with Christmas getting in the way of it being finalised. He will have perhaps the most accurate independent test data for the Cascade when it gets published.