抄袭了Arm的ComputeLibrary是否应该给出说明？

[izp001]

抄袭了Arm的ComputeLibrary是否应该给出说明？

[throneclay]

没人回复这个就关了，如果有问题可以随时跟我们联系

[yiicy]

你们开源是值得鼓励的，但是开源参考了就要说明，不然丢**人的脸

感谢指出问题，我们最开始的GEMM借鉴了compute library，现在sgemm.cpp里的接口已经弃用了，统一使用sgemm_prepacked.cpp里的接口，实现上也做了相应的变化，sgemm_prepacked.cpp的xblock将按照l2 cache大小划分

[yiicy]

既然开源了，就要遵守游戏规则

请问指的是哪部分文件

[izp001]

你们自己不知道吗？连注释都一样，有朋友问我了，我就给你指出来，你们最好修改以后再说没抄
sgemm_arm.cpp
/**

• Sgemm Kernel implementation for arm-v7a
• Assume that "Apanel" points to a chunk of A blocks (each size 6xK) in read-order.
• Assume that "Bpanel" points to a chunk of B blocks (each size 8xK) in read-order.
• Assume that "Cpanel" points to a chunk of C output blocks (each size 8x6),
• the chunks being arranged in a row major fashion.
• Note that the intent of this is that either ablocks or bblocks will be 1
• • this construction allows the output loop to proceed in either order.
    */
    而ComputeLibrary的generic.cpp
    ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/a32_sgemm_8x6/generic.cpp
    // Kernel implementation.
    //
    // Assume that "Apanel" points to a chunk of A blocks (each size 6xK) in read-order.
    // Assume that "Bpanel" points to a chunk of B blocks (each size 8xK) in read-order.
    // Assume that "Cpanel" points to a chunk of C output blocks (each size
    // 8x6), the chunks being arranged in a row major fashion.
    //
    // Note that the intent of this is that either ablocks or bblocks will be 1
    // - this construction allows the output loop to proceed in either order.

[izp001]

你们开源是值得鼓励的，但是开源参考了就要说明，不然丢**人的脸

[daquexian]

你们开源是值得鼓励的，但是开源参考了就要说明，不然丢**人的脸

感谢指出问题，我们最开始的GEMM借鉴了compute library，现在sgemm.cpp里的接口已经弃用了，统一使用sgemm_prepacked.cpp里的接口，实现上也做了相应的变化，sgemm_prepacked.cpp的xblock将按照l2 cache大小划分

这和弃不弃用无关。您能不能遵守开源协议、尊重原作者的权利呢？

[daquexian]

@throneclay

没人回复这个就关了，如果有问题可以随时跟我们联系

我联系了，你们的回复呢？

[xyoungli]

既然开源了，就要遵守游戏规则
您好，该问题与#531issue合并，如有疑问可以随时联系
#531