Re: [perf-discuss] Re: Arificial enlargement of binaries to enable VMPSS aks auto mpss

2006-03-16 Thread Eric Lowe
Matthew Allen wrote: I have an application that is suffering from high iTLB misses and I'd really like to enable VMPSS. I can create a binary with the text segment aligned on a 4M boundary, but the text is still mapped to 8K pages (as reported by pmap -xs). Is there anything I need to do to e

Re: [perf-discuss] Re: Arificial enlargement of binaries to enable VMPSS aks auto mpss

2006-03-16 Thread Bart Smaalders
Matthew Allen wrote: I have an application that is suffering from high iTLB misses and I'd really like to enable VMPSS. I can create a binary with the text segment aligned on a 4M boundary, but the text is still mapped to 8K pages (as reported by pmap -xs). Is there anything I need to do to e

[perf-discuss] Re: Arificial enlargement of binaries to enable VMPSS aks auto mpss

2006-03-16 Thread Matthew Allen
I have an application that is suffering from high iTLB misses and I'd really like to enable VMPSS. I can create a binary with the text segment aligned on a 4M boundary, but the text is still mapped to 8K pages (as reported by pmap -xs). Is there anything I need to do to enable large page suppor