I have never had so that makes personal experience a no. But I read a lot and also have theory as to why it will not work.
In all honesty with all my reading I read those that have saying the bad and disabling it and those who ask about. Almost no praises of it.
Let me throw out my theory of why it is flawed fundamentally.
L1<L2(<L3)<RAM<HDD, correct? See how the size keeps getting larger? Also speed gets slower either actual or in implementation. Why, how come? Most systems with Turbo, L1<L2(<L3)<RAM>=Turbo<HDD
Let me start with CPU cache as this is well documented and easily researchable. I believe those same concepts apply all the way up the chain but that is my own opinion can't find documentation to support.
Once L1 was on die it had a clear speed advantage over off die L2 or L3. Now that doesn't matter since all on die now. So why still keep multiple caches? Because The smaller L1 is faster because of it's smaller size vs L2, L3. L1 is important if the CPU makes a request and that is a hit on L1 that is the fastest. Intel and AMD decide the best size and they design different. Once they set the ideal size of L1. If you doubled the size of the L1 you would not double the the number hits you may only increase by 10%? And slow it down because of increase in size. Well they decide just not worth it. So they make L2 much larger (C2D) in my case L1 is 128KB L2 is 6MB's. Now that is going to give a very good shot at getting a hit if L1 does not but will not slow down L1 on what it would get hits on. Making L1 6MB's is slower real world than the L1/L2 I described. L3 is the exact same idea just a further extension.
So when you go beyond cache to RAM look at the great size increase? Even in the old days PIII say L2 512KB the RAM was easily 128MB/256MB a substantial increase in size. Current I have 6MB and 4GB RAM. That size increase is important as it gives a very good chance as all before if they did not provide a hit a good chance RAM will return a hit? When it doesn't you go to HDD and mine is 320GB and you will get a hit as nothing left to go to. But HDD hits are the slowest because of the HDD speed but also because all that was gone through prior takes time.
Some might say the speed of Cache is so fast how could it matter? It matters because of the incredible quantity of requests any latency has a cumulative affect. Even RAM which is fast suffers the same.
OK so how does any of this apply to Turbo? Because the jump from RAM to Turbo does not offer the increased size which as I tried to say is what gives a reasonable likely hood of getting a hit. That likely hood is what makes it worth going to that step. Turbo does not offer that. In most cases of recent you jumped to Turbo and it was smaller/same/at best double and that is not enough size increase to likely return a hit.
If you put 1GB RAM in and had 4GB Turbo I suspect adding Turbo would help but only because you starved the RAM. If you have 4GB RAM and don't get a hit there is not a beneficial statistical likely hood that what you need is in the next 4GB (Turbo) so wasting time going there and not finding retards performance since when you don't find you must go to HDD any way. If you go to Turbo 10 times and get 1 hit? Likely faster to have never gone there 9 times?
SSD's have further made Turbo theoretically misguided. It is the same technology/speed as Turbo but will get a hit so why have Turbo?
It is just flawed. Not that it started out that way. When first conceived I am sure it made sense. I understand what it might of offered and hope you all understand from my post how I mean that. But things changed and it lost any functional advantage. Intel being the giant they are, are slow to let go and further lose their investment without getting a return. But does not me you or I should underwrite their R&D.
Before someone points out HDD's have cache. Which seems at odds with my ever increasing size argument? It is not for several reasons. First some HDD cache is required for SATA. Second that cache is there to deal with the inner stand alone working of a HDD. It allows it to hold things and manage it operations to increase efficiency not so much that the cache will get the hit. It has also been demonstrated that going from 8MB to 16MB to 32MB cache has minimal impact on HDD performance. In most cases some if not all can be attributed to other improvements such as firmware, data density to name two.
I don't have to stick my hand in the fire to know I will get burned. I feel that way about Turbo.