I really am starting to become increasingly convinced that waterblock performance is hitting a wall.
With a ~100mm² CPU die size, and 6W of hydraulic pumping power, my waterblocks have exhibited roughly the following historical progression of effective convectional transfer efficiencies:
"Concept Block" - original WW precursor with plastic slit nozzle Prototype: ~62000 W/m²K
White Water x 1mm channels: ~67000 W/m²K
White Water x 0.8mm channels Prototype: ~71000 W/m²K
Free (un-nozzled) Jet Array Against flat base-plate (1mm jets) Prototype: ~55000 W/m²K
Mini-cupped free jet (un-nozzled) Prototype: ~65000 W/m²K
Cascade: ~72000 W/m²K
Cascade SS: ~74000 W/m²K
Cascade XXX Prototype: ~80000 W/m²K
Storm/G1 Prototype: ~50000 W/m²K
Storm/G3 Prototype: ~65000 W/m²K
Storm/G4: ~77000 W/m²K
Storm/G5: ~85000 W/m²K
Swiftech STORM (G4 Rev2): ~83000 W/m²K
Storm/G5 w/ G4 Rev2 optimisations (theoretical): ~91000 W/m²K (projected estimate)
Storm/G5 w/ G7-level optimisations (theoretical): ~94000 W/m²K (projected estimate)
Storm/G7: ~105000 W/m²K
Now in all that, the difference from the White Water to the Storm/G7 is about exactly 3.0C better for a 100W heat load on a 100mm² CPU die size. Overclocking-wise though the newer blocks do a fair bit better than what such a small temperature difference would otherwise imply.
The trek from 62000 to 105000 has been long and hard. In order to match a Storm/G7 in a copper implementation would take an h(eff) of around 120000 W/m²K, and to be honest I can't see that happening. Even if there was a block done in silver with an h(eff) of 120000 W/m²K, then performance would only improve by around 0.004 C/W (0.4C in 100W) [or by a projected ~0.25C better on the Procooling test charts].
We've seen micro-channel designs, pin grid array designs of various types, WW-variants by the dozen (the 1A-Cooling 1A-HV4 is effectively a WW-variant), and through it all we haven't seen a single design that threatens to hit anything much above 80000 W/m²K with ~6W of hydraulic pumping power on a ~100mm² CPU die.
Call it a magniloquent declaration if you will, and I'm sure plenty of you will
, but I'll wait to be proven wrong and I'll believe it when I see it independently tested, verified, and head-to-head against a G7 on a testbed that's measuring actual die temperatures (NOT IHS case-temps which I believe is fundamentally flawed due to uncontrollable variances in IHS-CPU die contact).