Re: [PATCH v3 0/2] arm/arm64: KVM: Optimize arm64 fp/simd, saves 30-50% on exits
On Thu, Jul 02, 2015 at 10:49:03AM -0700, Mario Smarduch wrote: On 07/01/2015 02:49 AM, Christoffer Dall wrote: On Wed, Jun 24, 2015 at 05:04:10PM -0700, Mario Smarduch wrote: Currently we save/restore fp/simd on each exit. Fist patch optimizes arm64 save/restore, we only do so on Guest access. hackbench and several lmbench tests show anywhere from 30% to above 50% optimzation achieved. In second patch 32-bit handler is updated to keep exit handling consistent with 64-bit code. 30-50% of what? The overhead or overall performance? Yes, so considering all exits to Host KVM anywhere from 30 to 50% didn't require an fp/simd switch. Anything else you like to see added here? No, I'm good with them. Marc is handling the tree these days so I'll leave it up to him if we want to adjust patch 1 or what to do. Thanks! -Christoffer -- To unsubscribe from this list: send the line unsubscribe kvm in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH v3 0/2] arm/arm64: KVM: Optimize arm64 fp/simd, saves 30-50% on exits
On 07/01/2015 02:49 AM, Christoffer Dall wrote: On Wed, Jun 24, 2015 at 05:04:10PM -0700, Mario Smarduch wrote: Currently we save/restore fp/simd on each exit. Fist patch optimizes arm64 save/restore, we only do so on Guest access. hackbench and several lmbench tests show anywhere from 30% to above 50% optimzation achieved. In second patch 32-bit handler is updated to keep exit handling consistent with 64-bit code. 30-50% of what? The overhead or overall performance? Yes, so considering all exits to Host KVM anywhere from 30 to 50% didn't require an fp/simd switch. Anything else you like to see added here? Changes since v1: - Addressed Marcs comments - Verified optimization improvements with lmbench and hackbench, updated commit message Changes since v2: - only for patch 2/2 - Reworked trapping to vfp access handler Changes since v3: - Only for patch 2/2 - Removed load_vcpu in switch_to_guest_vfp per Marcs comment - Got another chance to replace an unreferenced label with a comment Mario Smarduch (2): Optimize arm64 skip 30-50% vfp/simd save/restore on exits keep arm vfp/simd exit handling consistent with arm64 arch/arm/kvm/interrupts.S| 14 +++- arch/arm64/include/asm/kvm_arm.h |5 - arch/arm64/kvm/hyp.S | 46 +++--- 3 files changed, 55 insertions(+), 10 deletions(-) -- 1.7.9.5 -- To unsubscribe from this list: send the line unsubscribe kvm in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH v3 0/2] arm/arm64: KVM: Optimize arm64 fp/simd, saves 30-50% on exits
On Wed, Jun 24, 2015 at 05:04:10PM -0700, Mario Smarduch wrote: Currently we save/restore fp/simd on each exit. Fist patch optimizes arm64 save/restore, we only do so on Guest access. hackbench and several lmbench tests show anywhere from 30% to above 50% optimzation achieved. In second patch 32-bit handler is updated to keep exit handling consistent with 64-bit code. 30-50% of what? The overhead or overall performance? Changes since v1: - Addressed Marcs comments - Verified optimization improvements with lmbench and hackbench, updated commit message Changes since v2: - only for patch 2/2 - Reworked trapping to vfp access handler Changes since v3: - Only for patch 2/2 - Removed load_vcpu in switch_to_guest_vfp per Marcs comment - Got another chance to replace an unreferenced label with a comment Mario Smarduch (2): Optimize arm64 skip 30-50% vfp/simd save/restore on exits keep arm vfp/simd exit handling consistent with arm64 arch/arm/kvm/interrupts.S| 14 +++- arch/arm64/include/asm/kvm_arm.h |5 - arch/arm64/kvm/hyp.S | 46 +++--- 3 files changed, 55 insertions(+), 10 deletions(-) -- 1.7.9.5 -- To unsubscribe from this list: send the line unsubscribe kvm in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
[PATCH v3 0/2] arm/arm64: KVM: Optimize arm64 fp/simd, saves 30-50% on exits
Currently we save/restore fp/simd on each exit. Fist patch optimizes arm64 save/restore, we only do so on Guest access. hackbench and several lmbench tests show anywhere from 30% to above 50% optimzation achieved. In second patch 32-bit handler is updated to keep exit handling consistent with 64-bit code. Changes since v1: - Addressed Marcs comments - Verified optimization improvements with lmbench and hackbench, updated commit message Changes since v2: - only for patch 2/2 - Reworked trapping to vfp access handler Changes since v3: - Only for patch 2/2 - Removed load_vcpu in switch_to_guest_vfp per Marcs comment - Got another chance to replace an unreferenced label with a comment Mario Smarduch (2): Optimize arm64 skip 30-50% vfp/simd save/restore on exits keep arm vfp/simd exit handling consistent with arm64 arch/arm/kvm/interrupts.S| 14 +++- arch/arm64/include/asm/kvm_arm.h |5 - arch/arm64/kvm/hyp.S | 46 +++--- 3 files changed, 55 insertions(+), 10 deletions(-) -- 1.7.9.5 -- To unsubscribe from this list: send the line unsubscribe kvm in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html