So, after I checked up on the history yesterday, I'll look into the future today 
As my post yesterday showed, the performance of the fastest iMac about doubled (4185 to 8330) when the platform switched from C2D to iX.
What kind of performance increase can we expect, when the 2013 Mac Pro arrives?
SP Mac Pro
For the SP Mac Pro, the current W3565@3.2GHz has a PassMark score of 6027. We don't know which CPU will go into a new Mac Pro, but the E5-1650 might be a good guess with 6 cores ($583). Or an E5-1620 with only 4 cores ($294).
1620 has a PassMark score of 9673.
1650 has a PassMark score of 13242.
Based on this we would see an performance increase in the range of 60-120%, of course depending on how much of the PassMark score that translate into Geekbench points.
Now, this performance increase might seem ludicrous, but it is worth noting that the current maxed out iMac have a geekbench score of 11447 vs. the low-end Mac Pro which scores 9739. It is not unreasonable to expect the 2012 top iMac to have a GB of about 13-14000.
In other words, the 2013 low-end Mac Pro have to see an performance increase of about 44% just to match the 2012 high-end iMac. If a SP Mac Pro should maintain its relevancy, even the E5-1620 seem a bit on the low side. So, we'll probably see a 1650 which bumps the CPU performance compared to the current with 120% (and the GB score with about 100%?).
DP Mac Pro
For the DP Mac Pro, the E5645 might be replaced by 2630 and the current X5650 will have a successor in 2650.
PassMark scores:
E5645: 13931
X5650: 15528
2630: 19031
2650: 21661
So the DP Mac Pro will see a performance increase of about 37-40% it seems, based on CPU alone.
This is in same range as Anand reported: He experienced increases of 30-60%, probably because of other improvements found in the architecture.
If these numbers hold true, we could easily see a mid-range DP 2013 Mac Pro hit more than 30000 GB points.
Conclusion
As mentioned yesterday, a revolution is needed for Mac Pro. But with the CPUs most obvious to use, that revolution is kinda expected.
If we see the SP Mac Pro go to 19000 in GB points, while maintaining current pricing, it will $/GB-point be competitive with the strongest iMac again.
Bottom-line: There's so significant performance increases to be expected in the 2013 Mac Pros, that there's no reason whatsoever of purchasing a current-gen. Even purchasing a pre-gen and upgrading the CPU seems a poor proposition. Except, of course, that you need it right now. But then a Hackintosh might be a better way to hold over until next year.
As my post yesterday showed, the performance of the fastest iMac about doubled (4185 to 8330) when the platform switched from C2D to iX.
What kind of performance increase can we expect, when the 2013 Mac Pro arrives?
SP Mac Pro
For the SP Mac Pro, the current W3565@3.2GHz has a PassMark score of 6027. We don't know which CPU will go into a new Mac Pro, but the E5-1650 might be a good guess with 6 cores ($583). Or an E5-1620 with only 4 cores ($294).
1620 has a PassMark score of 9673.
1650 has a PassMark score of 13242.
Based on this we would see an performance increase in the range of 60-120%, of course depending on how much of the PassMark score that translate into Geekbench points.
Now, this performance increase might seem ludicrous, but it is worth noting that the current maxed out iMac have a geekbench score of 11447 vs. the low-end Mac Pro which scores 9739. It is not unreasonable to expect the 2012 top iMac to have a GB of about 13-14000.
In other words, the 2013 low-end Mac Pro have to see an performance increase of about 44% just to match the 2012 high-end iMac. If a SP Mac Pro should maintain its relevancy, even the E5-1620 seem a bit on the low side. So, we'll probably see a 1650 which bumps the CPU performance compared to the current with 120% (and the GB score with about 100%?).
DP Mac Pro
For the DP Mac Pro, the E5645 might be replaced by 2630 and the current X5650 will have a successor in 2650.
PassMark scores:
E5645: 13931
X5650: 15528
2630: 19031
2650: 21661
So the DP Mac Pro will see a performance increase of about 37-40% it seems, based on CPU alone.
This is in same range as Anand reported: He experienced increases of 30-60%, probably because of other improvements found in the architecture.
If these numbers hold true, we could easily see a mid-range DP 2013 Mac Pro hit more than 30000 GB points.
Conclusion
As mentioned yesterday, a revolution is needed for Mac Pro. But with the CPUs most obvious to use, that revolution is kinda expected.
If we see the SP Mac Pro go to 19000 in GB points, while maintaining current pricing, it will $/GB-point be competitive with the strongest iMac again.
Bottom-line: There's so significant performance increases to be expected in the 2013 Mac Pros, that there's no reason whatsoever of purchasing a current-gen. Even purchasing a pre-gen and upgrading the CPU seems a poor proposition. Except, of course, that you need it right now. But then a Hackintosh might be a better way to hold over until next year.
Last edited: