feature(lvgl_port): Initial support for ppa rendering in lvgl (BSP-563) #409

ThunderDai · 2024-10-15T04:03:18Z

  Mainly supports the following features:
   - color blend (simple fill)
   - color blend with opa 
   - blend normal (blend with argb8888) 
   - blend normal with color (color convert / memcpy)

This feature can only used in direct/full mode

CLAassistant · 2024-10-15T04:03:24Z

All committers have signed the CLA.

ThunderDai · 2024-10-16T01:59:34Z

@peter-marcisovsky @espzav hi, PTAL, this is my first commit for esp-bsp. Therefore, some code related rules are not familiar enough

peter-marcisovsky

Could you consider adding some tests to your feature? If you take a look at my #357 , there are functionality and benchmark unit tests. Would it be possible to do the same for this feature?

Also, can you take a look at the failing CI?

peter-marcisovsky · 2024-10-17T14:44:19Z

components/esp_lvgl_port/include/esp_lvgl_port_lv_blend.h

@@ -55,6 +55,21 @@ typedef struct {

 extern int lv_color_blend_to_argb8888_esp(asm_dsc_t *asm_dsc);

+// Just for compatibility with v9.2.0


I would not recommend this approach of solving compatibility issues with LVGL 9.2.
I tried to build lvlgl_port test app on your branch, with the HW accelerating enabled and the test app would not build, neither with LVGL9.1, nor with LVGL9.2.

peter@peter ➜ ~/esp/esp-bsp/components/esp_lvgl_port/test_apps/lvgl_port git:(pr/409) ✗ idf.py -DSDKCONFIG_DEFAULTS='sdkconfig.defaults;sdkconfig.ci.asm_render' set-target esp32s3 build

Is your feature working with LVGL9.1? If yes, could you consider using just LVGL9.1 in this MR?

We have agreed to make the HW acceleration support working with LVGL9.1, for now here, until we figure out what to do with this brekaing changes which was introduced in LVGL9.2.

ThunderDai · 2024-10-18T01:08:35Z

Could you consider adding some tests to your feature? If you take a look at my #357 , there are functionality and benchmark unit tests. Would it be possible to do the same for this feature?

Also, can you take a look at the failing CI?

yes, can add some test. I have already included the newly submitted lvgl_port component as part of my project and have run the benchmark without any issues. I will also try running the existing test_apps. It is worth noting that because PPA has implemented some of your current functions, it conflicts with the current header file. I have added a new header file myself. The following configuration is required:

CONFIG_LV_DRAW_SW_ASM_CUSTOM=y
CONFIG_LV_DRAW_SW_ASM_CUSTOM_INCLUDE="esp_lvgl_port_lv_blend_ppa.h"
CONFIG_LV_USE_CUSTOM_MALLOC=y

tore-espressif

@ThunderDai I'm sorry for the late start of review. I left few comments for start

tore-espressif · 2024-11-15T10:21:50Z

components/esp_lvgl_port/examples/mipi_dsi/main/idf_component.yml

+  espressif/esp32_p4_function_ev_board:
+    version: "*"
+    override_path: "../../../../../bsp/esp32_p4_function_ev_board"


For esp_lvgl_port, we try to avoid dependencies on any BSP. Can we removed it?

If not, we can place the example in different location,(not directly into esp_lvgl_port component)

tore-espressif · 2024-11-15T12:47:28Z

components/esp_lvgl_port/src/lvgl9/esp_lvgl_port_disp.c

@@ -286,7 +296,11 @@ static lv_display_t *lvgl_port_add_disp_priv(const lvgl_port_display_cfg_t *disp
        buff_caps |= MALLOC_CAP_DEFAULT;
    }

-    /* Use RGB internal buffers for avoid tearing effect */
+#if CONFIG_IDF_TARGET_ESP32P4


Suggested change

#if CONFIG_IDF_TARGET_ESP32P4

#if SOC_CACHE_INTERNAL_MEM_VIA_L1CACHE

Would this be better for forward compatibility?

tore-espressif · 2024-11-15T12:48:04Z

components/esp_lvgl_port/src/lvgl9/esp_lvgl_port_disp.c

-        if (buf1) {
-            free(buf1);
+        if (disp_ctx->draw_buffs[0]) {
+            free(disp_ctx->draw_buffs[0]);
        }
-        if (buf2) {
-            free(buf2);
+        if (disp_ctx->draw_buffs[1]) {
+            free(disp_ctx->draw_buffs[1]);


Thank you for this fix! It was already fixed in the master branch, so you might have to resolve some conflicts...

tore-espressif · 2024-11-15T12:50:14Z

components/esp_lvgl_port/CMakeLists.txt

@@ -77,12 +77,12 @@ if("usb_host_hid" IN_LIST build_components)
 endif()

 # Include SIMD assembly source code for rendering, only for (9.1.0 <= LVG_version < 9.2.0) and only for esp32 and esp32s3
-if((lvgl_ver VERSION_GREATER_EQUAL "9.1.0") AND (lvgl_ver VERSION_LESS "9.2.0"))
-    if(CONFIG_IDF_TARGET_ESP32 OR CONFIG_IDF_TARGET_ESP32S3)
+if((lvgl_ver VERSION_GREATER_EQUAL "9.1.0") AND (lvgl_ver VERSION_LESS_EQUAL "9.2.0"))


Does it really work with 9.2.0? we had some compatibility problems with this version. If yes, why can't support 9.2.2 too?

tore-espressif · 2024-11-18T14:28:24Z

components/esp_lvgl_port/test_apps/simd/main/lv_blend/include/lv_version.h

+#define LVGL_VERSION_MAJOR 9
+#define LVGL_VERSION_MINOR 1
+#define LVGL_VERSION_PATCH 0
+#define LVGL_VERSION_INFO ""


Why can't we use lv_version.h from the lvgl component?

tore-espressif · 2024-11-22T08:00:53Z

components/esp_lvgl_port/src/lvgl9/ppa/bsp_dma2d.c

+    dma2d_descriptor_t *tx_link_buffer = (dma2d_descriptor_t *)heap_caps_aligned_calloc(64, 1, 64, MALLOC_CAP_SPIRAM);
+    dma2d_descriptor_t *rx_link_buffer = (dma2d_descriptor_t *)heap_caps_aligned_calloc(64, 1, 64, MALLOC_CAP_SPIRAM);


The cache line size is configurable. Can be 64 or 128. Can we take the value from sdkconfig.h?

tore-espressif · 2024-11-22T08:01:48Z

components/esp_lvgl_port/src/lvgl9/ppa/bsp_dma2d.c

+    esp_cache_msync((void *)tx_dsc, 64, ESP_CACHE_MSYNC_FLAG_DIR_C2M | ESP_CACHE_MSYNC_FLAG_UNALIGNED);
+    esp_cache_msync((void *)rx_dsc, 64, ESP_CACHE_MSYNC_FLAG_DIR_C2M | ESP_CACHE_MSYNC_FLAG_UNALIGNED);


Same here, you assume cache line size == 64. But it can be 128 too.

Is the unaligned sync sage here? If yes, please provide better comment

espzav · 2025-03-27T15:00:11Z

I am not sure, why the benchmark not printed to comment, I checked in artifacts and it seems, that this optimization is not good.

Benchmark for BOARD esp32_p4_function_ev_board

DATE: 27.03.2025 14:06

LVGL version: 9.2.2

Name	Avg. CPU	Avg. FPS	Avg. time	render time	flush time
Empty screen	88% ₍₊₂₇₎	66 _(-21)	12 ₍₊₇₎	5	7 ₍₊₇₎
Moving wallpaper	99% ₍₊₇₎	22 _(-49)	40 ₍₊₃₀₎	38 ₍₊₃₁₎	2 _(-1)
Single rectangle	97% ₍₊₇₄₎	69 _(-20)	12 ₍₊₁₁₎	0 _(-1)	12 ₍₊₁₂₎
Multiple rectangles	97% ₍₊₅₃₎	69 _(-29)	11 ₍₊₅₎	7 ₍₊₃₎	4 ₍₊₂₎
Multiple RGB images	98% ₍₊₆₎	22 _(-29)	37 ₍₊₂₃₎	27 ₍₊₁₅₎	10 ₍₊₈₎
Multiple ARGB images	99% ₍₊₁₎	16 _(-8)	53 ₍₊₁₈₎	51 ₍₊₁₈₎	2
Rotated ARGB images	99%	3	280 ₍₊₃₃₎	267 ₍₊₂₂₎	13 ₍₊₁₁₎
Multiple labels	98% _(-1)	33 ₍₊₁₎	23 _(-3)	18 _(-6)	5 ₍₊₃₎
Screen sized text	10% ₍₊₃₎	89 ₍₊₁₎	8 ₍₊₄₎	2 _(-2)	6 ₍₊₆₎
Multiple arcs	90% _(-2)	34 _(-5)	23 ₍₊₁₎	19 _(-1)	4 ₍₊₂₎
Containers	99% ₍₊₄₃₎	22 _(-67)	34 ₍₊₂₉₎	28 ₍₊₂₃₎	6 ₍₊₆₎
Containers with overlay	99% ₍₊₂₎	13 _(-4)	67 ₍₊₁₆₎	57 ₍₊₉₎	10 ₍₊₇₎
Containers with opa	99% ₍₊₂₈₎	14 _(-62)	56 ₍₊₄₇₎	49 ₍₊₄₀₎	7 ₍₊₇₎
Containers with opa_layer	99%	8 _(-2)	117 ₍₊₃₄₎	109 ₍₊₂₈₎	8 ₍₊₆₎
Containers with scrolling	99%	12 _(-6)	72 ₍₊₂₂₎	68 ₍₊₂₀₎	4 ₍₊₂₎
Widgets demo	99%	16	51 _(-2)	43 _(-8)	8 ₍₊₆₎
All scenes avg.	91% ₍₊₁₅₎	31 _(-19)	55 ₍₊₁₇₎	49 ₍₊₁₂₎	6 ₍₊₅₎

ThunderDai · 2025-03-31T01:53:10Z

I am not sure, why the benchmark not printed to comment, I checked in artifacts and it seems, that this optimization is not good.

Benchmark for BOARD esp32_p4_function_ev_board

DATE: 27.03.2025 14:06

LVGL version: 9.2.2

Name Avg. CPU Avg. FPS Avg. time render time flush time
Empty screen 88% (+27) 66 (-21) 12 (+7) 5 7 (+7)
Moving wallpaper 99% (+7) 22 (-49) 40 (+30) 38 (+31) 2 (-1)
Single rectangle 97% (+74) 69 (-20) 12 (+11) 0 (-1) 12 (+12)
Multiple rectangles 97% (+53) 69 (-29) 11 (+5) 7 (+3) 4 (+2)
Multiple RGB images 98% (+6) 22 (-29) 37 (+23) 27 (+15) 10 (+8)
Multiple ARGB images 99% (+1) 16 (-8) 53 (+18) 51 (+18) 2
Rotated ARGB images 99% 3 280 (+33) 267 (+22) 13 (+11)
Multiple labels 98% (-1) 33 (+1) 23 (-3) 18 (-6) 5 (+3)
Screen sized text 10% (+3) 89 (+1) 8 (+4) 2 (-2) 6 (+6)
Multiple arcs 90% (-2) 34 (-5) 23 (+1) 19 (-1) 4 (+2)
Containers 99% (+43) 22 (-67) 34 (+29) 28 (+23) 6 (+6)
Containers with overlay 99% (+2) 13 (-4) 67 (+16) 57 (+9) 10 (+7)
Containers with opa 99% (+28) 14 (-62) 56 (+47) 49 (+40) 7 (+7)
Containers with opa_layer 99% 8 (-2) 117 (+34) 109 (+28) 8 (+6)
Containers with scrolling 99% 12 (-6) 72 (+22) 68 (+20) 4 (+2)
Widgets demo 99% 16 51 (-2) 43 (-8) 8 (+6)
All scenes avg. 91% (+15) 31 (-19) 55 (+17) 49 (+12) 6 (+5)

In fact, our PPA/DMA2D acceleration is currently only useful for specific scenarios:

Color Conversion (RGB565<-->RGB888)
Memory Copy
Data filling
Overlay transparent colors

And the above scenarios all need to have better effects on larger areas, while benchmarks are a comprehensive evaluation, and there are many small areas that need to be drawn separately.
If the above scenarios are evaluated separately, the acceleration effect is better, such as flashing the decoded JPEG data stream separately or filling the background color when sliding the menu.
In addition, it is recommended to enable tear resistance for all tests(CONFIG_BSP_DISPLAY_LVGL_AVOID_TEAR=y), otherwise it will not have much value in practical applications

ThunderDai · 2025-03-31T02:38:49Z

Benchmark Summary (9.2.2 ) --------- All scenes avg fps (based on avoid tear and direct mode)

color	using PPA	not use PPA
RGB565	31fps	23fps
RGB888	28fps(*25 fps)	19fps

From this table we can see, the avg fps can add about 8-9fps after using PPA/DMA2D

If we use PPA in RGB888, then will dsi underrun, so we need limit DMA2D speed by these code, and avg fps drop from 28 to 25fps, but still higher 6fps then not use ppa

#include "hal/axi_dma_ll.h"
#include "hal/axi_icm_ll.h"
#include "soc/mipi_dsi_bridge_struct.h"

int peak_level = 5; // 4 means 800MB/s, 5 means 400MB/s, 6 means 200MB/s
axi_icm_ll_set_qos_burstiness(AXI_ICM_MASTER_DMA2D, 256, AXI_ICM_ACCESS_WRITE);
axi_icm_ll_set_qos_peak_transaction_rate(AXI_ICM_MASTER_DMA2D, peak_level, peak_level + 1, AXI_ICM_ACCESS_WRITE);

ThunderDai · 2025-03-31T02:55:34Z

I am not sure, why the benchmark not printed to comment, I checked in artifacts and it seems, that this optimization is not good.

Benchmark for BOARD esp32_p4_function_ev_board

DATE: 27.03.2025 14:06

LVGL version: 9.2.2

Name Avg. CPU Avg. FPS Avg. time render time flush time
Empty screen 88% (+27) 66 (-21) 12 (+7) 5 7 (+7)
Moving wallpaper 99% (+7) 22 (-49) 40 (+30) 38 (+31) 2 (-1)
Single rectangle 97% (+74) 69 (-20) 12 (+11) 0 (-1) 12 (+12)
Multiple rectangles 97% (+53) 69 (-29) 11 (+5) 7 (+3) 4 (+2)
Multiple RGB images 98% (+6) 22 (-29) 37 (+23) 27 (+15) 10 (+8)
Multiple ARGB images 99% (+1) 16 (-8) 53 (+18) 51 (+18) 2
Rotated ARGB images 99% 3 280 (+33) 267 (+22) 13 (+11)
Multiple labels 98% (-1) 33 (+1) 23 (-3) 18 (-6) 5 (+3)
Screen sized text 10% (+3) 89 (+1) 8 (+4) 2 (-2) 6 (+6)
Multiple arcs 90% (-2) 34 (-5) 23 (+1) 19 (-1) 4 (+2)
Containers 99% (+43) 22 (-67) 34 (+29) 28 (+23) 6 (+6)
Containers with overlay 99% (+2) 13 (-4) 67 (+16) 57 (+9) 10 (+7)
Containers with opa 99% (+28) 14 (-62) 56 (+47) 49 (+40) 7 (+7)
Containers with opa_layer 99% 8 (-2) 117 (+34) 109 (+28) 8 (+6)
Containers with scrolling 99% 12 (-6) 72 (+22) 68 (+20) 4 (+2)
Widgets demo 99% 16 51 (-2) 43 (-8) 8 (+6)
All scenes avg. 91% (+15) 31 (-19) 55 (+17) 49 (+12) 6 (+5)

The defaults mode maybe not use avoid tear solution? so they can not compare

Mainly supports the following features: - color blend (simple fill) - color blend with opa - blend normal (blend with argb8888) - blend normal to color (color convert / memcpy)

github-actions bot changed the title ~~feature(lvgl_port): Initial support for ppa rendering in lvgl~~ feature(lvgl_port): Initial support for ppa rendering in lvgl (BSP-563) Oct 15, 2024

ThunderDai force-pushed the feat/add_ppa_for_lvgl_render branch from 0984d25 to 9723801 Compare October 15, 2024 06:26

ThunderDai marked this pull request as draft October 15, 2024 06:44

ThunderDai marked this pull request as ready for review October 15, 2024 07:34

peter-marcisovsky reviewed Oct 17, 2024

View reviewed changes

ThunderDai marked this pull request as draft October 21, 2024 08:47

ThunderDai force-pushed the feat/add_ppa_for_lvgl_render branch 2 times, most recently from 24e762a to 0b1609e Compare October 24, 2024 02:24

ThunderDai marked this pull request as ready for review October 24, 2024 02:26

ThunderDai force-pushed the feat/add_ppa_for_lvgl_render branch 6 times, most recently from 461a0f8 to 47b925e Compare October 28, 2024 06:32

tore-espressif reviewed Nov 15, 2024

View reviewed changes

tore-espressif reviewed Nov 22, 2024

View reviewed changes

mpolitowicz mentioned this pull request Mar 25, 2025

feat(lvgl_port): Used PPA for rotation on ESP32-P4 #352

Open

2 tasks

ThunderDai force-pushed the feat/add_ppa_for_lvgl_render branch 2 times, most recently from 9c82bb2 to b92dfa2 Compare March 27, 2025 09:44

espzav added the Run benchmark label Mar 27, 2025

ThunderDai force-pushed the feat/add_ppa_for_lvgl_render branch 2 times, most recently from c76daf9 to 7a10e47 Compare March 27, 2025 12:42

ThunderDai force-pushed the feat/add_ppa_for_lvgl_render branch 2 times, most recently from db70cfb to 642ef56 Compare March 31, 2025 11:57

feat(lvgl_port): Initial support for ppa rendering in lvgl

cb78c54

Mainly supports the following features: - color blend (simple fill) - color blend with opa - blend normal (blend with argb8888) - blend normal to color (color convert / memcpy)

ThunderDai force-pushed the feat/add_ppa_for_lvgl_render branch from 642ef56 to cb78c54 Compare April 8, 2025 03:34

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feature(lvgl_port): Initial support for ppa rendering in lvgl (BSP-563) #409

feature(lvgl_port): Initial support for ppa rendering in lvgl (BSP-563) #409

ThunderDai commented Oct 15, 2024 •

edited

Loading

CLAassistant commented Oct 15, 2024 •

edited

Loading

ThunderDai commented Oct 16, 2024

peter-marcisovsky left a comment •

edited

Loading

peter-marcisovsky Oct 17, 2024

ThunderDai commented Oct 18, 2024

tore-espressif left a comment

tore-espressif Nov 15, 2024

tore-espressif Nov 15, 2024

tore-espressif Nov 15, 2024

tore-espressif Nov 15, 2024

tore-espressif Nov 18, 2024

tore-espressif Nov 22, 2024

tore-espressif Nov 22, 2024

espzav commented Mar 27, 2025

ThunderDai commented Mar 31, 2025 •

edited

Loading

Benchmark for BOARD esp32_p4_function_ev_board

ThunderDai commented Mar 31, 2025 •

edited

Loading

ThunderDai commented Mar 31, 2025

Benchmark for BOARD esp32_p4_function_ev_board

		@@ -55,6 +55,21 @@ typedef struct {

		extern int lv_color_blend_to_argb8888_esp(asm_dsc_t *asm_dsc);

		// Just for compatibility with v9.2.0

	#if CONFIG_IDF_TARGET_ESP32P4
	#if SOC_CACHE_INTERNAL_MEM_VIA_L1CACHE

		dma2d_descriptor_t tx_link_buffer = (dma2d_descriptor_t )heap_caps_aligned_calloc(64, 1, 64, MALLOC_CAP_SPIRAM);
		dma2d_descriptor_t rx_link_buffer = (dma2d_descriptor_t )heap_caps_aligned_calloc(64, 1, 64, MALLOC_CAP_SPIRAM);

		esp_cache_msync((void *)tx_dsc, 64, ESP_CACHE_MSYNC_FLAG_DIR_C2M \| ESP_CACHE_MSYNC_FLAG_UNALIGNED);
		esp_cache_msync((void *)rx_dsc, 64, ESP_CACHE_MSYNC_FLAG_DIR_C2M \| ESP_CACHE_MSYNC_FLAG_UNALIGNED);

feature(lvgl_port): Initial support for ppa rendering in lvgl (BSP-563) #409

Are you sure you want to change the base?

feature(lvgl_port): Initial support for ppa rendering in lvgl (BSP-563) #409

Conversation

ThunderDai commented Oct 15, 2024 • edited Loading

CLAassistant commented Oct 15, 2024 • edited Loading

ThunderDai commented Oct 16, 2024

peter-marcisovsky left a comment • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ThunderDai commented Oct 18, 2024

tore-espressif left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

espzav commented Mar 27, 2025

Benchmark for BOARD esp32_p4_function_ev_board

ThunderDai commented Mar 31, 2025 • edited Loading

Benchmark for BOARD esp32_p4_function_ev_board

ThunderDai commented Mar 31, 2025 • edited Loading

ThunderDai commented Mar 31, 2025

Benchmark for BOARD esp32_p4_function_ev_board

ThunderDai commented Oct 15, 2024 •

edited

Loading

CLAassistant commented Oct 15, 2024 •

edited

Loading

peter-marcisovsky left a comment •

edited

Loading

ThunderDai commented Mar 31, 2025 •

edited

Loading

ThunderDai commented Mar 31, 2025 •

edited

Loading