checkasm: hevc: Iterate over features first, then over bitdepths

Message ID 1467147569-92141-1-git-send-email-martin@martin.st
State Committed
Commit 67cb2c0f73ec08bdcecd675c1ffe25c3a5b26ef2
Headers show

Commit Message

Martin Storsjö June 28, 2016, 8:59 p.m.
This avoids listing the same feature multiple times in the
test output. Previously the output contained something like this:

SSE2:
 - hevc_mc.qpel              [OK]
 - hevc_mc.epel              [OK]
 - hevc_mc.unweighted_pred   [OK]
 - hevc_mc.qpel              [OK]
 - hevc_mc.epel              [OK]
 - hevc_mc.unweighted_pred   [OK]
---
 tests/checkasm/hevc_mc.c | 18 +++++++++++++-----
 1 file changed, 13 insertions(+), 5 deletions(-)

Comments

Luca Barbato June 28, 2016, 9:13 p.m. | #1
On 28/06/16 22:59, Martin Storsjö wrote:
> This avoids listing the same feature multiple times in the
> test output. Previously the output contained something like this:
> 

Seems fine.
Diego Biurrun June 29, 2016, 10:35 a.m. | #2
On Tue, Jun 28, 2016 at 11:59:29PM +0300, Martin Storsjö wrote:
> This avoids listing the same feature multiple times in the
> test output. Previously the output contained something like this:
> 
> SSE2:
>  - hevc_mc.qpel              [OK]
>  - hevc_mc.epel              [OK]
>  - hevc_mc.unweighted_pred   [OK]
>  - hevc_mc.qpel              [OK]
>  - hevc_mc.epel              [OK]
>  - hevc_mc.unweighted_pred   [OK]

Maybe just print the tested bit depth along with the component instead?

Diego
Martin Storsjö June 29, 2016, 6:11 p.m. | #3
On Wed, 29 Jun 2016, Diego Biurrun wrote:

> On Tue, Jun 28, 2016 at 11:59:29PM +0300, Martin Storsjö wrote:
>> This avoids listing the same feature multiple times in the
>> test output. Previously the output contained something like this:
>> 
>> SSE2:
>>  - hevc_mc.qpel              [OK]
>>  - hevc_mc.epel              [OK]
>>  - hevc_mc.unweighted_pred   [OK]
>>  - hevc_mc.qpel              [OK]
>>  - hevc_mc.epel              [OK]
>>  - hevc_mc.unweighted_pred   [OK]
>
> Maybe just print the tested bit depth along with the component instead?

As concluded on irc, not really useful - will push.

// Martin

Patch

diff --git a/tests/checkasm/hevc_mc.c b/tests/checkasm/hevc_mc.c
index 7139369..385c33e 100644
--- a/tests/checkasm/hevc_mc.c
+++ b/tests/checkasm/hevc_mc.c
@@ -304,17 +304,25 @@  void checkasm_check_hevc_mc(void)
 
     for (bit_depth = 8; bit_depth <= 10; bit_depth++) {
         ff_hevc_dsp_init(&h, bit_depth);
-
         check_qpel(&h, buf16_0, buf16_1, buf8_0, mcbuffer, bit_depth);
-        report("qpel");
+    }
+    report("qpel");
 
+    for (bit_depth = 8; bit_depth <= 10; bit_depth++) {
+        ff_hevc_dsp_init(&h, bit_depth);
         check_epel(&h, buf16_0, buf16_1, buf8_0, mcbuffer, bit_depth);
-        report("epel");
+    }
+    report("epel");
 
+    for (bit_depth = 8; bit_depth <= 10; bit_depth++) {
+        ff_hevc_dsp_init(&h, bit_depth);
         check_unweighted_pred(&h, buf8_0, buf8_1, buf16_0, buf16_1, bit_depth);
-        report("unweighted_pred");
+    }
+    report("unweighted_pred");
 
+    for (bit_depth = 8; bit_depth <= 10; bit_depth++) {
+        ff_hevc_dsp_init(&h, bit_depth);
         check_weighted_pred(&h, buf8_0, buf8_1, buf16_0, buf16_1, bit_depth);
-        report("weighted_pred");
     }
+    report("weighted_pred");
 }