From: "Martin Storsjö" <martin@martin.st>
To: FFmpeg development discussions and patches <ffmpeg-devel@ffmpeg.org>
Cc: Niklas Haas <git@haasn.dev>
Subject: Re: [FFmpeg-devel] [PATCH v3 04/17] tests/checkasm: generalize DEF_CHECKASM_CHECK_FUNC to floats
Date: Tue, 27 May 2025 11:24:26 +0300 (EEST)
Message-ID: <739d1310-7b32-b97f-d533-f1536ea659a@martin.st> (raw)
In-Reply-To: <20250527081242.22892-5-ffmpeg@haasn.xyz>
On Tue, 27 May 2025, Niklas Haas wrote:
> From: Niklas Haas <git@haasn.dev>
>
> We split the standard macro into its body (implementation) and declaration,
> and use a macro argument in place of the raw `memcmp` call, with the major
> difference that we now take the number of pixels to compare instead of the
> number of bytes (to match the signature of float_near_ulp_array).
> ---
> tests/checkasm/checkasm.c | 52 ++++++++++++++++++++++++++-------------
> tests/checkasm/checkasm.h | 7 ++++++
> 2 files changed, 42 insertions(+), 17 deletions(-)
>
> diff --git a/tests/checkasm/checkasm.c b/tests/checkasm/checkasm.c
> index 71d1e5766c..f393a0cb96 100644
> --- a/tests/checkasm/checkasm.c
> +++ b/tests/checkasm/checkasm.c
> @@ -1187,14 +1187,8 @@ static int check_err(const char *file, int line,
> return 0;
> }
>
> -#define DEF_CHECKASM_CHECK_FUNC(type, fmt) \
> -int checkasm_check_##type(const char *file, int line, \
> - const type *buf1, ptrdiff_t stride1, \
> - const type *buf2, ptrdiff_t stride2, \
> - int w, int h, const char *name, \
> - int align_w, int align_h, \
> - int padding) \
> -{ \
> +#define DEF_CHECKASM_CHECK_BODY(compare, type, fmt) \
> +do { \
> int64_t aligned_w = (w - 1LL + align_w) & ~(align_w - 1); \
> int64_t aligned_h = (h - 1LL + align_h) & ~(align_h - 1); \
> int err = 0; \
> @@ -1204,7 +1198,7 @@ int checkasm_check_##type(const char *file, int line, \
> stride1 /= sizeof(*buf1); \
> stride2 /= sizeof(*buf2); \
> for (y = 0; y < h; y++) \
> - if (memcmp(&buf1[y*stride1], &buf2[y*stride2], w*sizeof(*buf1))) \
> + if (!compare(&buf1[y*stride1], &buf2[y*stride2], w)) \
> break; \
> if (y != h) { \
> if (check_err(file, line, name, w, h, &err)) \
> @@ -1226,38 +1220,50 @@ int checkasm_check_##type(const char *file, int line, \
> buf2 -= h*stride2; \
> } \
> for (y = -padding; y < 0; y++) \
> - if (memcmp(&buf1[y*stride1 - padding], &buf2[y*stride2 - padding], \
> - (w + 2*padding)*sizeof(*buf1))) { \
> + if (!compare(&buf1[y*stride1 - padding], &buf2[y*stride2 - padding], \
> + w + 2*padding)) { \
> if (check_err(file, line, name, w, h, &err)) \
> return 1; \
> fprintf(stderr, " overwrite above\n"); \
> break; \
> } \
> for (y = aligned_h; y < aligned_h + padding; y++) \
> - if (memcmp(&buf1[y*stride1 - padding], &buf2[y*stride2 - padding], \
> - (w + 2*padding)*sizeof(*buf1))) { \
> + if (!compare(&buf1[y*stride1 - padding], &buf2[y*stride2 - padding], \
> + w + 2*padding)) { \
> if (check_err(file, line, name, w, h, &err)) \
> return 1; \
> fprintf(stderr, " overwrite below\n"); \
> break; \
> } \
> for (y = 0; y < h; y++) \
> - if (memcmp(&buf1[y*stride1 - padding], &buf2[y*stride2 - padding], \
> - padding*sizeof(*buf1))) { \
> + if (!compare(&buf1[y*stride1 - padding], &buf2[y*stride2 - padding], \
> + padding)) { \
> if (check_err(file, line, name, w, h, &err)) \
> return 1; \
> fprintf(stderr, " overwrite left\n"); \
> break; \
> } \
> for (y = 0; y < h; y++) \
> - if (memcmp(&buf1[y*stride1 + aligned_w], &buf2[y*stride2 + aligned_w], \
> - padding*sizeof(*buf1))) { \
> + if (!compare(&buf1[y*stride1 + aligned_w], &buf2[y*stride2 + aligned_w], \
> + padding)) { \
> if (check_err(file, line, name, w, h, &err)) \
> return 1; \
> fprintf(stderr, " overwrite right\n"); \
> break; \
> } \
> return err; \
> +} while (0)
> +
> +#define cmp_int(a, b, len) (!memcmp(a, b, (len) * sizeof(*(a))))
> +#define DEF_CHECKASM_CHECK_FUNC(type, fmt) \
> +int checkasm_check_##type(const char *file, int line, \
> + const type *buf1, ptrdiff_t stride1, \
> + const type *buf2, ptrdiff_t stride2, \
> + int w, int h, const char *name, \
> + int align_w, int align_h, \
> + int padding) \
> +{ \
> + DEF_CHECKASM_CHECK_BODY(cmp_int, type, fmt); \
> }
>
> DEF_CHECKASM_CHECK_FUNC(uint8_t, "%02x")
> @@ -1265,3 +1271,15 @@ DEF_CHECKASM_CHECK_FUNC(uint16_t, "%04x")
> DEF_CHECKASM_CHECK_FUNC(uint32_t, "%08x")
> DEF_CHECKASM_CHECK_FUNC(int16_t, "%6d")
> DEF_CHECKASM_CHECK_FUNC(int32_t, "%9d")
> +
> +int checkasm_check_float_ulp(const char *file, int line,
> + const float *buf1, ptrdiff_t stride1,
> + const float *buf2, ptrdiff_t stride2,
> + int w, int h, const char *name,
> + unsigned max_ulp, int align_w, int align_h,
> + int padding)
> +{
> + #define cmp_float(a, b, len) float_near_ulp_array(a, b, max_ulp, len)
> + DEF_CHECKASM_CHECK_BODY(cmp_float, float, "%g");
> + #undef cmp_float
> +}
> diff --git a/tests/checkasm/checkasm.h b/tests/checkasm/checkasm.h
> index ad7ed10613..ec01bd6207 100644
> --- a/tests/checkasm/checkasm.h
> +++ b/tests/checkasm/checkasm.h
> @@ -423,6 +423,13 @@ DECL_CHECKASM_CHECK_FUNC(uint32_t);
> DECL_CHECKASM_CHECK_FUNC(int16_t);
> DECL_CHECKASM_CHECK_FUNC(int32_t);
>
> +int checkasm_check_float_ulp(const char *file, int line,
> + const float *buf1, ptrdiff_t stride1,
> + const float *buf2, ptrdiff_t stride2,
> + int w, int h, const char *name,
> + unsigned max_elp, int align_w, int align_h,
Typo - max_ulp?
Other than that, thanks, this looks reasonable!
// Martin
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
next prev parent reply other threads:[~2025-05-27 8:24 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-05-27 7:55 [FFmpeg-devel] (no subject) Niklas Haas
2025-05-27 7:55 ` [FFmpeg-devel] [PATCH v3 01/17] swscale/format: rename legacy format conversion table Niklas Haas
2025-05-27 7:55 ` [FFmpeg-devel] [PATCH v3 02/17] swscale/format: add ff_fmt_clear() Niklas Haas
2025-05-27 7:55 ` [FFmpeg-devel] [PATCH v3 03/17] tests/checkasm: increase number of runs in between measurements Niklas Haas
2025-05-27 7:55 ` [FFmpeg-devel] [PATCH v3 04/17] tests/checkasm: generalize DEF_CHECKASM_CHECK_FUNC to floats Niklas Haas
2025-05-27 8:24 ` Martin Storsjö [this message]
2025-05-27 7:55 ` [FFmpeg-devel] [PATCH v3 05/17] swscale: add SWS_UNSTABLE flag Niklas Haas
2025-05-27 7:55 ` [FFmpeg-devel] [PATCH v3 06/17] swscale/ops: introduce new low level framework Niklas Haas
2025-05-27 7:55 ` [FFmpeg-devel] [PATCH v3 07/17] swscale/optimizer: add high-level ops optimizer Niklas Haas
2025-05-27 7:55 ` [FFmpeg-devel] [PATCH v3 08/17] swscale/ops_internal: add internal ops backend API Niklas Haas
2025-05-27 7:55 ` [FFmpeg-devel] [PATCH v3 09/17] swscale/ops: add dispatch layer Niklas Haas
2025-05-27 7:55 ` [FFmpeg-devel] [PATCH v3 10/17] swscale/optimizer: add packed shuffle solver Niklas Haas
2025-05-27 7:55 ` [FFmpeg-devel] [PATCH v3 11/17] swscale/ops_chain: add internal abstraction for kernel linking Niklas Haas
2025-05-27 7:55 ` [FFmpeg-devel] [PATCH v3 12/17] swscale/ops_backend: add reference backend basend on C templates Niklas Haas
2025-05-27 7:55 ` [FFmpeg-devel] [PATCH v3 13/17] swscale/ops_memcpy: add 'memcpy' backend for plane->plane copies Niklas Haas
2025-05-27 7:55 ` [FFmpeg-devel] [PATCH v3 14/17] swscale/x86: add SIMD backend Niklas Haas
2025-05-30 2:23 ` Michael Niedermayer
2025-05-30 10:34 ` Niklas Haas
2025-05-27 7:55 ` [FFmpeg-devel] [PATCH v3 15/17] tests/checkasm: add checkasm tests for swscale ops Niklas Haas
2025-05-27 8:25 ` Martin Storsjö
2025-05-27 7:55 ` [FFmpeg-devel] [PATCH v3 16/17] swscale/format: add new format decode/encode logic Niklas Haas
2025-05-27 7:55 ` [FFmpeg-devel] [PATCH v3 17/17] swscale/graph: allow experimental use of new format handler Niklas Haas
2025-05-27 8:29 ` [FFmpeg-devel] (no subject) Kieran Kunhya via ffmpeg-devel
2025-05-27 8:51 ` Niklas Haas
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=739d1310-7b32-b97f-d533-f1536ea659a@martin.st \
--to=martin@martin.st \
--cc=ffmpeg-devel@ffmpeg.org \
--cc=git@haasn.dev \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
This inbox may be cloned and mirrored by anyone:
git clone --mirror https://master.gitmailbox.com/ffmpegdev/0 ffmpegdev/git/0.git
# If you have public-inbox 1.1+ installed, you may
# initialize and index your mirror using the following commands:
public-inbox-init -V2 ffmpegdev ffmpegdev/ https://master.gitmailbox.com/ffmpegdev \
ffmpegdev@gitmailbox.com
public-inbox-index ffmpegdev
Example config snippet for mirrors.
AGPL code for this site: git clone https://public-inbox.org/public-inbox.git