I would like to mention that the response Ian posted was/is a nice read.
I am not sure what to write in response. Let me just note that adding
formalisms to Mesa would lead to both higher performance and higher
safety. If you do not want C++ in Mesa, maybe it would be acceptable
for Mesa developer
Hi,
On 18.10.2016 20:12, Jan Ziak wrote:
[...]
Never profile with -O0 or disabled function inlining.
Seriously?
Nobody's going to take seriously optimization results taken from
non-optimized builds.
Mesa uses -g -O2
with --enable-debug, so that's what you should use too. Don't use any
o
Hi Jan,
On 18.10.2016 00:07, Jan Ziak wrote:
This patch replaces the ir_variable_refcount_entry's linked-list
with an array-list.
The array-list has local storage which does not require ANY additional
allocations if the list has small number of elements. The size of this
storage is configurabl
On 10/18/2016 10:12 AM, Jan Ziak wrote:
>> Regarding C++ templates, the compiler doesn't use them. If u_vector
>> (Dave Airlie?) provides the same functionality as your array, I
>> suggest we use u_vector instead.
>
> Let me repeat what you just wrote, because it is unbelievable: You are
> advisin
On Tue, Oct 18, 2016 at 8:04 PM, Marek Olšák wrote:
> On Tue, Oct 18, 2016 at 7:12 PM, Jan Ziak <0xe2.0x9a.0...@gmail.com>
> wrote:
> >> Regarding C++ templates, the compiler doesn't use them. If u_vector
> >> (Dave Airlie?) provides the same functionality as your array, I
> >> suggest we use u_v
Perf stat results for shader-db:
This is measured on an AMD Kaveri CPU.
gcc-6.2.0 -fno-omit-frame-pointer -g -O2
Unpatched:
$ cd shader-db
$ ../run-upstream perfstat-u --repeat=5 -- ./run -1 shaders >/dev/null
Performance counter stats for './run -1 shaders' (5 runs):
13689.962374
On Tue, Oct 18, 2016 at 7:12 PM, Jan Ziak <0xe2.0x9a.0...@gmail.com> wrote:
>> Regarding C++ templates, the compiler doesn't use them. If u_vector
>> (Dave Airlie?) provides the same functionality as your array, I
>> suggest we use u_vector instead.
>
> Let me repeat what you just wrote, because it
> Regarding C++ templates, the compiler doesn't use them. If u_vector
> (Dave Airlie?) provides the same functionality as your array, I
> suggest we use u_vector instead.
Let me repeat what you just wrote, because it is unbelievable: You are
advising the use of non-templated collection types in C+
On Tue, Oct 18, 2016 at 3:55 PM, Eero Tamminen
wrote:
> Hi,
>
> On 18.10.2016 16:25, Jan Ziak wrote:
>>
>> On Tue, Oct 18, 2016 at 3:12 PM, Nicolai Hähnle
>> wrote:
>>>
>>> On 18.10.2016 15:07, Jan Ziak wrote:
On Tue Oct 18 09:29:59 UTC 2016, Eero Tamminen wrote:
>
> On 18.10.20
Hi,
On 18.10.2016 16:25, Jan Ziak wrote:
On Tue, Oct 18, 2016 at 3:12 PM, Nicolai Hähnle wrote:
On 18.10.2016 15:07, Jan Ziak wrote:
On Tue Oct 18 09:29:59 UTC 2016, Eero Tamminen wrote:
On 18.10.2016 01:07, Jan Ziak wrote:
- The total number of executed instructions goes down from 64.184 t
Hi Michael,
thanks for the suggestions about code formatting. I formatted the
whole fast_list.h file in the Netbeans editor and uploaded a new
revision of the patch.
Jan
On Tue, Oct 18, 2016 at 1:42 PM, Michael Schellenberger Costa
wrote:
> Hi Jan,
>
> On 18.10.2016 00:07, Jan Ziak wrote:
>>
>>
On Tue, Oct 18, 2016 at 3:32 PM, Emil Velikov wrote:
> On 17 October 2016 at 23:07, Jan Ziak <0xe2.0x9a.0...@gmail.com> wrote:
>
>> - exec_list assign_list;
>> + arraylist assign_list;
>>
> Just an FYI - when people started using C++ for glsl there was an
> agreement that templating is not to
On 17 October 2016 at 23:07, Jan Ziak <0xe2.0x9a.0...@gmail.com> wrote:
> - exec_list assign_list;
> + arraylist assign_list;
>
Just an FYI - when people started using C++ for glsl there was an
agreement that templating is not to be used.
Not sure if it still holds true, nowadays.
Emil
__
This patch replaces the ir_variable_refcount_entry's linked-list
with an array-list.
The array-list has local storage which does not require ANY additional
allocations if the list has small number of elements. The size of this
storage is configurable for each variable.
Benchmark results for "./ru
On Tue, Oct 18, 2016 at 3:12 PM, Nicolai Hähnle wrote:
> On 18.10.2016 15:07, Jan Ziak wrote:
>>
>> On Tue Oct 18 09:29:59 UTC 2016, Eero Tamminen wrote:
>>>
>>> On 18.10.2016 01:07, Jan Ziak wrote:
- The total number of executed instructions goes down from 64.184 to
63.797
g
On 18.10.2016 15:07, Jan Ziak wrote:
On Tue Oct 18 09:29:59 UTC 2016, Eero Tamminen wrote:
On 18.10.2016 01:07, Jan Ziak wrote:
- The total number of executed instructions goes down from 64.184 to 63.797
giga-instructions when Mesa is compiled with "gcc -O0 ..."
Please don't do performance
On Tue Oct 18 09:29:59 UTC 2016, Eero Tamminen wrote:
> On 18.10.2016 01:07, Jan Ziak wrote:
>> - The total number of executed instructions goes down from 64.184 to 63.797
>> giga-instructions when Mesa is compiled with "gcc -O0 ..."
>
>Please don't do performance related decisions based on data
Hi,
On 18.10.2016 01:07, Jan Ziak wrote:
This patch replaces the ir_variable_refcount_entry's linked-list
with an array-list.
The array-list has local storage which does not require ANY additional
allocations if the list has small number of elements. The size of this
storage is configurable for
On 18 October 2016 at 08:45, Thomas Helland wrote:
>
> 18. okt. 2016 00.07 skrev "Jan Ziak" <0xe2.0x9a.0...@gmail.com>:
>>
>> This patch replaces the ir_variable_refcount_entry's linked-list
>> with an array-list.
>>
>> The array-list has local storage which does not require ANY additional
>> allo
On Tue, Oct 18, 2016 at 12:45 AM, Thomas Helland
wrote:
> I can't quite tell, as Gmail tends to mangle whitespace stuff,
> but it looks like there might be some style issues with
> not everything following the three-space indent, no tabs
> policy that mesa tries to stick to.
I followed the 3-spac
18. okt. 2016 00.07 skrev "Jan Ziak" <0xe2.0x9a.0...@gmail.com>:
>
> This patch replaces the ir_variable_refcount_entry's linked-list
> with an array-list.
>
> The array-list has local storage which does not require ANY additional
> allocations if the list has small number of elements. The size of
This patch replaces the ir_variable_refcount_entry's linked-list
with an array-list.
The array-list has local storage which does not require ANY additional
allocations if the list has small number of elements. The size of this
storage is configurable for each variable.
Benchmark results for "./ru
22 matches
Mail list logo
