• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Major performance boost for AMD CPUs touted....

Hi, cat,

Looks very interesting, I have downloaded it but will research before applying to Vishera.

andy.
 
No one is using the x87 instruction set any more so I can't see this making a big difference, most if not all modern software/games will be using SSE or better.
 
I'm more than a bit out of the loop these days when it comes to CPUs but isn't AMD's lagging behind in superPi more due to the fact their CPUs lack the ability of more recent Intel ones to prioritise and boost performance on a smaller number of cores when the workload is dependant on a smaller number of threads rather than it being poor support for an outdated set of instructions? which should mean that Kaveri based CPUs with its ability to prioritise workload in a similiar fashion should see a massive jump in performance in SuperPi and gaming.
 
Nutella33 said:
would think its because superpi uses a defunct instruction set. no point wasting time and money on it just because some benchers cling to it.
Click to expand...

It might be out of date but any performance boost anywhere down the line can't be a bad thing for AMD, would it cost that much to implement it or leave in, would it not help else where?

As l'm 58 and know very little at all compared to some, l'm not knocking AMD but every little help's. Always glad to see when the forum run's a benchmark thread and some one post's result's from a AMD CPU more power to them..
 
OLDPHART said:
It might be out of date but any performance boost anywhere down the line can't be a bad thing for AMD, would it cost that much to implement it or leave in, would it not help else where?

As l'm 58 and know very little at all compared to some, l'm not knocking AMD but every little help's. Always glad to see when the forum run's a benchmark thread and some one post's result's from a AMD CPU more power to them..
Click to expand...

but its not a performance boost in anything apart from one program which only some benchers use occasionally. there is no real world outcome or benefit so why would you use a team to increase performance in a benchmark that came out in 1995. As i have read it, the boost will only apply to that program.

AMD do need to improve their single threaded performance tho but then that has come from the design choices they have made.

should mean that Kaveri based CPUs with its ability to prioritise workload in a similiar fashion should see a massive jump in performance in SuperPi and gaming.
Click to expand...
i think its mainly the instruction set. other benchmarks show AMD's lack of single threaded performance but its quite a bit less drastic than the usual superpi result comparison.
 
Last edited:
It supposedly improves X87 performance an X87 is not used by much modern software. OTH,it seems the engine in Skyrim uses X87 instructions,since it is based on a very old one,so it quite possibly improve performance. It might have been patched with newer patches though,so I can't say for sure!
 
isnt some of the physx code for CPU compiled in x87 in version 2.x.x? perhaps theres a slight improvement there as well if i have understood what the patch is about correctly.
 
Rroff said:
I'm more than a bit out of the loop these days when it comes to CPUs but isn't AMD's lagging behind in superPi more due to the fact their CPUs lack the ability of more recent Intel ones to prioritise and boost performance on a smaller number of cores when the workload is dependant on a smaller number of threads rather than it being poor support for an outdated set of instructions? which should mean that Kaveri based CPUs with its ability to prioritise workload in a similiar fashion should see a massive jump in performance in SuperPi and gaming.
Click to expand...

Q6600 @ 4Ghz vs a Phenom II 955BE there's a 4 minute difference so 25-30% in favour of the Q6600

A Q6600 is not 25% faster than a Phenom II clock for clock in anything but superPi.
 
OK well I applied the patch just for lols. It may show improvements in superpi, have not tested it and won't, but does not make any noticeable difference in the CPU generally which is why I tested it. Probably not worth doing it for typical gaming or other taskload.

It required a bios reset as well, something to consider.

The CPU seems unaffected by patching so it is probably correct that the x87 code is largely redundant in todays programmes.
 
Dunno if I'd call a Q6600 faster.
Perhaps IPC wise, it's minor.

The Q6600 won't clock as high, and then you've got the CPU-NB gains, the Phenom II is definitely better.
 
Martini1991 said:
Dunno if I'd call a Q6600 faster.
Perhaps IPC wise, it's minor.

The Q6600 won't clock as high, and then you've got the CPU-NB gains, the Phenom II is definitely better.
Click to expand...

Depends what your doing, if I can find my old post on it the Q6600 can put a 965BE to shame in a lot of stuff. Old CPUs tho now.
 
You must log in or register to reply here.
Back
Top Bottom