Anywhere from a 5% to a 30% drop in performance
Even with that type of drop, it would still be a nice improvement to move to Intel x64 chips. We are continuing to use "Itaniums" to run our OE software on the HP-UX platform (Intel Itanium Processors with speeds from 1.73 GHz to 2.53 GHz).
Depending on the workload, and the number of Kernel calls that get done, it could be worse:
The impact of this change will vary wildly depending on workload. Applications that are heavily dependent on user programs and which don't call into the kernel often will see very little impact; games, for example, should see very little change. But applications that call into the operating system extensively, typically to perform disk or network operations, can see a much more substantial impact. In synthetic benchmarks that do nothing but make kernel calls, the difference can be substantial, dropping from five million kernel calls per second to two-to-three million.
It would be interesting if Progress runs the ATM benchmark tool against these types of patches/fixes, and if so, could they disclose the results of before/after the fix?
Great suggestion for gus bjorklund gus
james, you mean you want me to do actual work?
we will have to see how this pans out. there are two separate issues (Meltdown and Spectre) that are not related. and the effects of the patches will vary depending on the specific X86 processor type and model.
In case you are interested, the meltdown issue can be summarized as follows:
suppose you have code that does the following:
if (some boolean expression) then
There are two possible outcomes, depending on the result of the boolean expression.
Modern intel processors do what is called “speculative execution”. Simply put, this means they can execute both branches of the if at the same time as the expression is being evaluated and then throw away the result that is not needed. A modern Intel chip can have as many as 150 instructions (or maybe more) in flight in the instruction pipeline.
The defect is that full privilege checks are not always executed on all the memory references. This makes it possible for one process (or thread) to access memory locations that should not be visible. While this has been known for some years amongst cpu designers, it was thought that the rate at which wrong memory could be accessed was so low that it was not practical to exploit. They were very wrong.
AMD processors are not susceptible to Meltdown because Intel patented their particular speculative execution technique and security flaw and would not license it to AMD because it gave Intel a performance advantage. So AMD has to do full privilege checks on all branches.