Re: [PATCH] lookup_object: split up displacement penalty for hash collisions

[Stefan Beller <stefanbeller@googlemail.com>]

[-- Attachment #1: Type: text/plain, Size: 1922 bytes --]

On 08/16/2013 11:26 AM, Thomas Rast wrote:
> Thomas Rast <trast@inf.ethz.ch> writes:
> 
>> Stefan Beller <stefanbeller@googlemail.com> writes:
>>
>>> However please do check if this patch brings the promised performance
>>> on your own, as you're likely using different hardware and another
>>> software setup. Feel free to share your performance differences.
>>
>> I get this on an i7-M620 laptop from t/perf/p0001-rev-list.sh:
>>
>>   Test                               HEAD                next                    
>>   -------------------------------------------------------------------------------
>>   0001.1: rev-list --all             6.29(6.03+0.22)     6.33(6.06+0.24) +0.6%   
>>   0001.2: rev-list --all --objects   53.22(52.48+0.54)   54.90(54.15+0.55) +3.2%*
>>   -------------------------------------------------------------------------------
>>   Significance hints:  '.' 0.1  '*' 0.05  '**' 0.01  '***' 0.001
> [...]
>> I trust the laptop numbers less because it has far more thermal (and
>> thus throttling) issues, but the runs do show a significant difference,
>> though less than you claimed.
> 
> Well, as I feared... another run on the same laptop:
> 
> Test                               HEAD                next                                            
> ------------------------------------------------------------------------------
> 0001.1: rev-list --all             6.41(6.14+0.24)     6.36(6.10+0.23) -0.9%* 
> 0001.2: rev-list --all --objects   54.60(53.84+0.55)   54.23(53.50+0.53) -0.7%
> ------------------------------------------------------------------------------
> Significance hints:  '.' 0.1  '*' 0.05  '**' 0.01  '***' 0.001
> 

I did some more tests as well, and I seem to have just been lucky with
the results initially posted. Now I got a negative impact as well on one
test, so that patch is not worth for includsion.

Thanks,
Stefan


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 899 bytes --]