diff --git a/18-demonstrate_NICE_VNICE_acceleration_of_the_Nuclei_Model_through_profiling.md b/18-demonstrate_NICE_VNICE_acceleration_of_the_Nuclei_Model_through_profiling.md index a54083f..3e26e30 100644 --- a/18-demonstrate_NICE_VNICE_acceleration_of_the_Nuclei_Model_through_profiling.md +++ b/18-demonstrate_NICE_VNICE_acceleration_of_the_Nuclei_Model_through_profiling.md @@ -1,6 +1,6 @@ # 通过Profiling展示Nuclei Model NICE/VNICE指令加速 -> 由于 Nuclei Model 仅支持Linux版本,所以此文档的测试都是基于 Nuclei Studio 的 Linux版本 (>= 2024.06) 完成的。 +> Nuclei Model 已支持 Windows/Linux 版本,此文档测试都是基于 Nuclei Studio 的 Windows 版本 (>= 2025.10) 完成的。 ## 背景描述 @@ -15,7 +15,7 @@ Nuclei Model Profiling 的优势: - model 中内建了 gprof 功能,无需 Profiling 库和 `gcc -pg` 选项就可以产生 Profiling 文件 - 采取了指令级别的采样,可以进行指令级别的 Profiling 分析 -在[NucleiStudio_User_Guide.pdf](https://download.nucleisys.com/upload/files/doc/nucleistudio/Nuclei_Studio_User_Guide.202406.pdf)相关章节对 Nuclei Model 如何仿真性能分析配置已经有较详细的描述,此文档以一个例子来展示其实际应用。 +在[NucleiStudio_User_Guide.pdf](https://download.nucleisys.com/upload/files/doc/nucleistudio/Nuclei_Studio_User_Guide.202510.pdf)相关章节对 Nuclei Model 如何仿真性能分析配置已经有较详细的描述,此文档以一个例子来展示其实际应用。 ### NICE/VNICE 自定义指令加速 @@ -28,7 +28,7 @@ Nuclei Model Profiling 的优势: ### 环境准备 -Nuclei Studio:[NucleiStudio 2024.06 Linux](https://download.nucleisys.com/upload/files/nucleistudio/NucleiStudio_IDE_202406-lin64.tgz) +Nuclei Studio:[NucleiStudio 2025.10 Windows](https://download.nucleisys.com/upload/files/nucleistudio/NucleiStudio_IDE_202510-win64.zip) ### Model Profiling @@ -67,18 +67,23 @@ __RV_CSR_SET(CSR_MSTATUS, MSTATUS_XS); 首先将 `aes_debug.h` 中的 `LOCAL_DEBUG` 打开,准备测试 AES 算法的整体 cycle 数。 -Nuclei Model 仿真程序需要配置 Nuclei Studio 中的 RVProf 运行配置,打开 Nuclei Studio 主菜单栏的 `Run` 选项的 `Run Configurations` 后,先在 `Main` 选项卡中选择编译好的 elf 文件路径,然后在 `RVProf` 选项卡 -的 `Config options` 中完成 model 运行配置 `--trace=1 --gprof=1 --logdir=Debug`,`--trace=1` 表示开启 rvtrace,`--gprof=1` 表示开启 gprof 功能生成 `*.gmon` 文件,`--logdir=Debug` 则表示最终生成的 `*.rvtrace` 文件、`*.gmon` 文件存存放的路径为当前工程下的 Debug 目录,取消勾选 `Start RVProf locally`,然后点击 `Apply` 和 `Run`,model 就开始运行程序了。 +Model 仿真程序需要配置 Nuclei Studio 中的 `GDB Nuclei Model riscv Debugging` 配置项,步骤如下: -![image-Main_configuration](asserts/images/18/Main_configuration.png) +1. 打开 Nuclei Studio 主菜单栏的 `Run` 选项的 `Run Configurations` +2. 选择 `GDB Nuclei Model riscv Debugging` 配置项,右键选择 `New Configuration`,会自动生成项目名的 Model 配置页面,launch bar也会同步更新 +3. 在右侧 `Main` 选项卡中点击 `Search Project...` 选择编译好的 elf 文件 +4. 在右侧 `Debugger` 选项卡中选择 `Browse` 找到 Nuclei Model 可执行程序默认路径:`NucleiStudio/toolchain/nucleimodel/bin/xl_cpumodel.exe` +5. 在右侧 `Debugger` 选项卡中的 `Nuclei Setup` 中完成 model 运行配置, 选择 `Nuclei RISC-V Core` 和 `Other Extensions` 需要保持和 `Nuclei Settings` 的 `Core` 和 `Other extensions` 配置一致,`Other Extensions` 为空时不传递此参数, `Enable Nuclei Model RVTrace` 表示运行时生成 rvtrace,`More options` 加上 `--gprof=1` 开启 Profiling 功能, 然后点击 `Apply` 和 `Run`,model 就开始运行程序了 -![image-RVProf_configuration](asserts/images/18/RVProf_configuration.png) + ![image-Ori_Project_Model_Config](asserts/images/18/Ori_Project_Model_Config.png) -在 Console 中会看到 `Total elapsed time` 说明 model 已经完成仿真了,得到 AES 算法整体消耗 154988 cycle。 +> Nuclei Studio (< 2025.10) 只能使用 `Run Configurations` 中的 `Nuclei Model` 来配置 model,Nuclei Studio (>= 2025.10) 建议切换到使用 `GDB Nuclei Model riscv Debugging` 来配置 + +在 Console 中会看到 `Total elapsed real time` 说明 model 已经完成仿真了,得到 AES 算法整体消耗 161108 cycle。 ![image-ase_demo_cycle](asserts/images/18/aes_demo_cycle.png) -将 `aes_debug.h` 中的 `LOCAL_DEBUG` 关掉去掉程序打印,为了准确测试 Profiling 数据,确保 Nuclei Studio 的 launch bar 为 `aes_demo Debug`, 重新 Run model: +将 `aes_debug.h` 中的 `LOCAL_DEBUG` 关掉去掉程序打印,为了准确测试 Profiling 数据,确保 Nuclei Studio 的 launch bar 为 `aes_demo Debug`, 重新 Run model,运行结束后会生成 Profiling 文件: ![image-aes_demo_profiling](asserts/images/18/aes_demo_profiling.png) @@ -279,12 +284,14 @@ static void aes_mix_columns_enc( #### step6:在 Nuclei Model 中实现 NICE/VNICE 指令 -首先需要下载支持用户配置自定义 **NICE/VNICE** 指令的原始 Nuclei Model 软件包[原始model软件包下载](https://drive.weixin.qq.com/s?k=ABcAKgdSAFccHEiAsM),解压软件包为 `xlmodel_nice`,然后将其导入 Nuclei Studio。 +首先需要下载支持用户配置自定义 **NICE/VNICE** 指令的原始 Nuclei Model 软件包[原始model软件包下载](https://drive.weixin.qq.com/s?k=ABcAKgdSAFcCirwEWY),解压软件包为 `xlmodel_nice`,然后将其导入 Nuclei Studio。 导入步骤:File->Import->Projects from Folder or Archive->Next->Directory->选择 `xlmodel_nice`->Finish即可 如何使用 Nuclei Model 以及查看 `xlmodel_nice` 软件包的目录结构可以参考[Nuclei Model介绍](https://doc.nucleisys.com/nuclei_tools/xlmodel/),`xlmodel_nice` 是由CMake构建的,用户无需修改即可编译,在 -编译前选择 Nuclei Studio 的 launch bar 的 `xlmodel_nice`,然后点击编译,确保软件包本身编译通过,编译生成的 elf 文件所在路径为 `build/default/xl_cpumodel`: +编译前选择 Nuclei Studio 的 launch bar 的 `xlmodel_nice`,然后点击编译,确保软件包本身编译通过: + +> Nuclei Studio (< 2025.10) 生成的 elf 文件所在路径为 `build/default/xl_cpumodel` ![image-xlmodel_nice_build](asserts/images/18/xlmodel_nice_build.png) @@ -307,26 +314,22 @@ AES demo 中定义的 **NICE/VNICE** 指令实现如下图,通过指令的 `op 以上介绍了用户如何从原始 Nuclei Model 软件包添加自定义 **NICE/VNICE** 指令,接下来需要将新编译出的 model 可执行程序导入到 Nuclei Studio 中,为了不和 Nuclei Studio 原始 model 名称混淆,可以将 model 导入到 `NucleiStudio/toolchain/nucleimodel/bin_aes/` 的创建路径下,我们提供了两种 model 可执行程序获取方式: -1. 实现 AES demo **NICE/VNICE** 指令的 Nuclei model 软件包[添加AES NICE指令model软件包](https://drive.weixin.qq.com/s?k=ABcAKgdSAFc1wrUKu1),编译后将 `xl_cpumodel` 可执行程序导入上述路径。 -2. 编译好的 model 的可执行程序 [xl_cpumodel](https://drive.weixin.qq.com/s?k=ABcAKgdSAFc03skCLp),直接导入上述路径。 +1. 实现 AES demo **NICE/VNICE** 指令的 Nuclei model 软件包[添加AES NICE指令model软件包](https://drive.weixin.qq.com/s?k=ABcAKgdSAFcrcrb4T6),编译后将 `xl_cpumodel` 可执行程序导入上述路径。 +2. 编译好的 model 的可执行程序 [xl_cpumodel](https://drive.weixin.qq.com/s?k=ABcAKgdSAFcB1NbrL1),直接导入上述路径。 #### step7:热点函数再分析 **注意:** 请务必完成 step6 中介绍的实现了 **NICE/VNICE** 指令的 model 导入 Nuclei Studio 中才能用 model Run `aes_demo_nice` 工程。 -首先打开 Nuclei Studio 主菜单栏的 `Run` 选项的 `Run Configurations`,model 配置需要重新添加一份 Nuclei Studio 中的 RVProf 运行配置 `aes_demo_nice Debug`: - -![image-new_rvprof_configuration](asserts/images/18/new_rvprof_configuration.png) - -将 `Main` 选项卡的 `Project` 通过 `Browse` 改为 `aes_demo_nice`,`C/C++ Application` 通过 `Search Project` 改为 `aes_demo_nice.elf`: +首先打开 Nuclei Studio 主菜单栏的 `Run` 选项的 `Run Configurations`,model 配置需要重新添加新的 `GDB Nuclei Model riscv Debugging` 运行配置 `aes_demo_nice Debug`,在 `Main` 选项卡中选择 `aes_demo_nice.elf`: -![image-aes_main_configuration](asserts/images/18/aes_main_configuration.png) +![image-aes_demo_nice_model_config_main](asserts/images/18/aes_demo_nice_model_config_main.png) -然后将 `RVProf` 选项卡中的 model 执行路径 `Executable path` 改为 step6 中新修改 model 的执行路径: `.../NucleiStudio/toolchain/nucleimodel/bin_aes/xl_cpumodel`: +然后在 `Debugger` 选项卡中的 model 执行路径 `Executable path` 改为 step6 中新修改 model 的执行路径: `.../NucleiStudio/toolchain/nucleimodel/bin_aes/xl_cpumodel`: -![image-aes_rvprof_configuration](asserts/images/18/aes_rvprof_configuration.png) +![image-aes_demo_nice_model_config_debugger](asserts/images/18/aes_demo_nice_model_config_debugger.png) -运行前将 `aes_debug.h` 中的 `LOCAL_DEBUG` 打开,测试优化后 AES 算法的整体 cycle 数,选择 Nuclei Studio 的 launch bar 的 `aes_demo_nice Debug`后 Run model,得到 AES 算法优化后整体消耗 cycle 数从优化前的 154988 降到了 35619 cycle。 +运行前将 `aes_debug.h` 中的 `LOCAL_DEBUG` 打开,测试优化后 AES 算法的整体 cycle 数,选择 Nuclei Studio 的 launch bar 的 `aes_demo_nice Debug`后 Run model,得到 AES 算法优化后整体消耗 cycle 数从优化前的 161108 降到了 42066 cycle。 ![image-aes_demo_nice_cycle](asserts/images/18/aes_demo_nice_cycle.png) @@ -349,7 +352,7 @@ AES demo 中定义的 **NICE/VNICE** 指令实现如下图,通过指令的 `op | AES Program Total | Before Optimization | NICE/VNICE Optimization | |-----------------------------|---------------------|--------------------------| -| Cycles | 154,988 | 35,619 | +| Cycles | 161,108 | 42,066 | AES加解密 NICE/VNICE demo:[优化后AES工程链接下载](https://drive.weixin.qq.com/s?k=ABcAKgdSAFc5f6zPQW) diff --git a/19-rapid_verification_of_NICE_VNICE_acceleration_with_Nuclei_Model_and_NICE_Wizard.md b/19-rapid_verification_of_NICE_VNICE_acceleration_with_Nuclei_Model_and_NICE_Wizard.md index 07bc31b..d872992 100644 --- a/19-rapid_verification_of_NICE_VNICE_acceleration_with_Nuclei_Model_and_NICE_Wizard.md +++ b/19-rapid_verification_of_NICE_VNICE_acceleration_with_Nuclei_Model_and_NICE_Wizard.md @@ -1,6 +1,6 @@ # Nuclei Model结合Nice Wizard快速验证NICE/VNICE指令加速 -> Nuclei Model 已支持 Windows/Linux 版本,此文档测试都是基于 Nuclei Studio 的 Windows 版本 (>= 2025.02) 完成的。 +> Nuclei Model 已支持 Windows/Linux 版本,此文档测试都是基于 Nuclei Studio 的 Windows 版本 (>= 2025.10) 完成的。 ## 背景描述 @@ -27,13 +27,13 @@ Nuclei NICE Wizard 是 Nuclei Studio 上提供的 `NICE/VNICE` 指令生成控 **Nuclei Studio**: -- [NucleiStudio 202502 Windows](https://download.nucleisys.com/upload/files/nucleistudio/NucleiStudio_IDE_202502-win64.zip) -- [NucleiStudio 202502 Linux](https://download.nucleisys.com/upload/files/nucleistudio/NucleiStudio_IDE_202502-lin64.tgz) +- [NucleiStudio 202510 Windows](https://download.nucleisys.com/upload/files/nucleistudio/NucleiStudio_IDE_202510-win64.zip) +- [NucleiStudio 202510 Linux](https://download.nucleisys.com/upload/files/nucleistudio/NucleiStudio_IDE_202510-lin64.tgz) **xlmodel_nice**: -- [原始`xlmodel_nice`软件包 Windows](https://drive.weixin.qq.com/s?k=ABcAKgdSAFcYbanf6c) -- [原始`xlmodel_nice`软件包 Linux](https://drive.weixin.qq.com/s?k=ABcAKgdSAFczrqzDI1) +- [原始`xlmodel_nice`软件包 Windows](https://drive.weixin.qq.com/s?k=ABcAKgdSAFcCirwEWY) +- [原始`xlmodel_nice`软件包 Linux](https://drive.weixin.qq.com/s?k=ABcAKgdSAFcTgr1Dbv) ### Nuclei Model运行原始程序 @@ -55,17 +55,19 @@ Nuclei NICE Wizard 是 Nuclei Studio 上提供的 `NICE/VNICE` 指令生成控 ![image-Ori_Project_Nuclei_Settting](asserts/images/19/Ori_Project_Nuclei_Settting.png) -Model 仿真程序需要配置 Nuclei Studio 中的 `Nuclei Model` 配置项,步骤如下: +Model 仿真程序需要配置 Nuclei Studio 中的 `GDB Nuclei Model riscv Debugging` 配置项,步骤如下: 1. 打开 Nuclei Studio 主菜单栏的 `Run` 选项的 `Run Configurations` -2. 选择 `Nuclei Model` 配置项,右键选择 `New Configuration`,会自动生成项目名的 Model 配置页面,launch bar也会同步更新 +2. 选择 `GDB Nuclei Model riscv Debugging` 配置项,右键选择 `New Configuration`,会自动生成项目名的 Model 配置页面,launch bar也会同步更新 3. 在右侧 `Main` 选项卡中点击 `Search Project...` 选择编译好的 elf 文件 -4. 在右侧 `Model` 选项卡中选择 `Browse` 找到 Nuclei Model 可执行程序默认路径:`NucleiStudio/toolchain/nucleimodel/bin/xl_cpumodel.exe` -5. 在右侧 `Model` 选项卡中的 `Config options` 中完成 model 运行配置: `--cpu=n900fd --trace=1 --logdir=Debug`, `--cpu` 和 `--ext` 需要保持和 `Nuclei Settings` 的 `Core` 和 `Other extensions` 配置一致,`--ext` 为空时不传递此参数,`--trace=1` 表示开启 rvtrace,`--logdir=Debug` 则表示最终生成的 `*.rvtrace` 文件存存放路径为当前工程下的 Debug 目录,然后点击 `Apply` 和 `Run`,model 就开始运行程序了 +4. 在右侧 `Debugger` 选项卡中选择 `Browse` 找到 Nuclei Model 可执行程序默认路径:`NucleiStudio/toolchain/nucleimodel/bin/xl_cpumodel.exe` +5. 在右侧 `Debugger` 选项卡中的 `Nuclei Setup` 中完成 model 运行配置, 选择 `Nuclei RISC-V Core` 和 `Other Extensions` 需要保持和 `Nuclei Settings` 的 `Core` 和 `Other extensions` 配置一致,`Other Extensions` 为空时不传递此参数, `Enable Nuclei Model RVTrace` 表示运行时生成 rvtrace,然后点击 `Apply` 和 `Run`,model 就开始运行程序了 ![image-Ori_Project_Model_Config](asserts/images/19/Ori_Project_Model_Config.png) -在 Console 中会看到 `Total elapsed time` 说明 model 已经完成仿真了,程序会提取标量矩阵乘算法函数 `golden_case` 的执行指令数和 cycle 数如下: +> Nuclei Studio (< 2025.10) 只能使用 `Run Configurations` 中的 `Nuclei Model` 来配置 model,Nuclei Studio (>= 2025.10) 建议切换到使用 `GDB Nuclei Model riscv Debugging` 来配置 + +在 Console 中会看到 `Total elapsed real time` 说明 model 已经完成仿真了,程序会提取标量矩阵乘算法函数 `golden_case` 的执行指令数和 cycle 数如下: ![image-Ori_Project_Model_Run](asserts/images/19/Ori_Project_Model_Run.png) @@ -79,6 +81,8 @@ Model 仿真程序需要配置 Nuclei Studio 中的 `Nuclei Model` 配置项, 在编译 `xlmodel_nice` 前需先配置好 xlmodel 的编译环境 ([xlmodel_nice 编译环境配置](https://doc.nucleisys.com/nuclei_tools/xlmodel/intro.html#nice-build)),然后编译确保原始软件包可以成功编译生成 model 的可执行程序: +> Nuclei Studio (< 2025.10) 生成的 elf 文件所在路径为 `build/default/xl_cpumodel` + ![image-Ori_Model_Nice_Build](asserts/images/19/Ori_Model_Nice_Build.png) **step2:NICE Wizard生成NICE指令替换** @@ -196,13 +200,13 @@ for (int32_t kk = 0; kk < 4; kk++) | instret/cycle | golden_case | nice_case | vnice_case | golden / nice | golden / vnice | nice / vnice | |-----------------------------|---------------------|--------------------------|---------------------------|---------------------------|---------------------------|---------------------------| | instret | 2854 | 730 | 88 | 3.91 | 32.43 | 8.30 | -| cycle | 3859 | 964 | 121 | 4.00 | 31.89 | 7.97 | +| cycle | 3844 | 964 | 122 | 3.99 | 31.51 | 7.90 | 用户通过研究现有算法的优化策略,就可以将构想快速通过 NICE Wizard 生成相关 `NICE/VNICE` 指令,再通过 Nuclei Studio 导入 `xlmodel_nice` 软件包实现指令,编写应用程序指令优化 case,就可以很快的利用 Nuclei Model 验证算法优化效果,整个测试过程只需使用 Nuclei Studio 就可以完成。 [优化后的工程下载链接](https://drive.weixin.qq.com/s?k=ABcAKgdSAFc0dskAJG) -[优化后的`xlmodel_nice`软件包](https://drive.weixin.qq.com/s?k=ABcAKgdSAFcZ14pFy7) +[优化后的`xlmodel_nice`软件包](https://drive.weixin.qq.com/s?k=ABcAKgdSAFcbA9mEgt) diff --git a/README.md b/README.md index e65a2f7..df2a370 100644 --- a/README.md +++ b/README.md @@ -30,7 +30,7 @@ Click [this link](https://doc.nucleisys.com/nuclei_studio_supply/) to see online ## Documents -> Generated by `python3 update.py` @ 2025-09-26 09:48:46 +> Generated by `python3 update.py` @ 2025-12-05 16:17:56 - [1. 因内存不足,导致在Nuclei Studio中启动qemu失败](1-cannot-setup-guestmemory.md) - [2. windows 11下使用Nuclei Studio进行qemu调试程序时报错](2-qemu-glib-gio-unexpectedly.md) diff --git a/asserts/images/18/Ori_Project_Model_Config.png b/asserts/images/18/Ori_Project_Model_Config.png new file mode 100755 index 0000000..c9a57c8 Binary files /dev/null and b/asserts/images/18/Ori_Project_Model_Config.png differ diff --git a/asserts/images/18/aes_demo_cycle.png b/asserts/images/18/aes_demo_cycle.png index b3e86b5..f506404 100755 Binary files a/asserts/images/18/aes_demo_cycle.png and b/asserts/images/18/aes_demo_cycle.png differ diff --git a/asserts/images/18/aes_demo_nice_cycle.png b/asserts/images/18/aes_demo_nice_cycle.png index 1f615da..39e0604 100755 Binary files a/asserts/images/18/aes_demo_nice_cycle.png and b/asserts/images/18/aes_demo_nice_cycle.png differ diff --git a/asserts/images/18/aes_demo_nice_model_config_debugger.png b/asserts/images/18/aes_demo_nice_model_config_debugger.png new file mode 100755 index 0000000..706058b Binary files /dev/null and b/asserts/images/18/aes_demo_nice_model_config_debugger.png differ diff --git a/asserts/images/18/aes_demo_nice_model_config_main.png b/asserts/images/18/aes_demo_nice_model_config_main.png new file mode 100755 index 0000000..c331dd1 Binary files /dev/null and b/asserts/images/18/aes_demo_nice_model_config_main.png differ diff --git a/asserts/images/18/aes_demo_profiling.png b/asserts/images/18/aes_demo_profiling.png index 0a6ad97..c66ea33 100755 Binary files a/asserts/images/18/aes_demo_profiling.png and b/asserts/images/18/aes_demo_profiling.png differ diff --git a/asserts/images/18/nice_inst_intro.png b/asserts/images/18/nice_inst_intro.png index 632d86d..c9581cf 100755 Binary files a/asserts/images/18/nice_inst_intro.png and b/asserts/images/18/nice_inst_intro.png differ diff --git a/asserts/images/18/xlmodel_nice_aes.png b/asserts/images/18/xlmodel_nice_aes.png index 7f2da6f..0b468ee 100755 Binary files a/asserts/images/18/xlmodel_nice_aes.png and b/asserts/images/18/xlmodel_nice_aes.png differ diff --git a/asserts/images/18/xlmodel_nice_build.png b/asserts/images/18/xlmodel_nice_build.png index 08467ac..50ecb17 100755 Binary files a/asserts/images/18/xlmodel_nice_build.png and b/asserts/images/18/xlmodel_nice_build.png differ diff --git a/asserts/images/19/Import_xlmodel_nice.png b/asserts/images/19/Import_xlmodel_nice.png index 2a87440..be9d89b 100755 Binary files a/asserts/images/19/Import_xlmodel_nice.png and b/asserts/images/19/Import_xlmodel_nice.png differ diff --git a/asserts/images/19/NICE_Project_Model_Config.png b/asserts/images/19/NICE_Project_Model_Config.png index 669f84c..5f712b0 100755 Binary files a/asserts/images/19/NICE_Project_Model_Config.png and b/asserts/images/19/NICE_Project_Model_Config.png differ diff --git a/asserts/images/19/NICE_Project_Model_Run.png b/asserts/images/19/NICE_Project_Model_Run.png index 1c00d00..456a3ff 100755 Binary files a/asserts/images/19/NICE_Project_Model_Run.png and b/asserts/images/19/NICE_Project_Model_Run.png differ diff --git a/asserts/images/19/Ori_Model_Nice_Build.png b/asserts/images/19/Ori_Model_Nice_Build.png index b6f6af6..2ff9562 100755 Binary files a/asserts/images/19/Ori_Model_Nice_Build.png and b/asserts/images/19/Ori_Model_Nice_Build.png differ diff --git a/asserts/images/19/Ori_Project_Model_Config.png b/asserts/images/19/Ori_Project_Model_Config.png index e8fe9fa..d8debcb 100755 Binary files a/asserts/images/19/Ori_Project_Model_Config.png and b/asserts/images/19/Ori_Project_Model_Config.png differ diff --git a/asserts/images/19/Ori_Project_Model_Run.png b/asserts/images/19/Ori_Project_Model_Run.png index 523dfd1..6613314 100755 Binary files a/asserts/images/19/Ori_Project_Model_Run.png and b/asserts/images/19/Ori_Project_Model_Run.png differ diff --git a/asserts/images/19/VNICE_Project_Model_Config.png b/asserts/images/19/VNICE_Project_Model_Config.png index f51fba4..ef9377f 100755 Binary files a/asserts/images/19/VNICE_Project_Model_Config.png and b/asserts/images/19/VNICE_Project_Model_Config.png differ diff --git a/asserts/images/19/VNICE_Project_Model_Run.png b/asserts/images/19/VNICE_Project_Model_Run.png index 739c94e..38e7df4 100755 Binary files a/asserts/images/19/VNICE_Project_Model_Run.png and b/asserts/images/19/VNICE_Project_Model_Run.png differ