Only Notes for Translators Krita Manual Contribution Guide For first timers General philosophy Protocol Other Images for the Manual Tools for making screenshots The appropriate file format for the job figure caption. Needless to say, don’t submit images that cannot be licensed under either license. Protocol So here we line out all the boring workflows. Tagging and Branches Adding and removing text will ARM? Please remember, that for better performance it is still recommended to make sure that all buffers are aligned to SIMD-word boundary (that is, 32 bytes for AVX and 16 bytes for SSE). Fixed-size chunks

ARM? Please remember, that for better performance it is still recommended to make sure that all buffers are aligned to SIMD-word boundary (that is, 32 bytes for AVX and 16 bytes for SSE). Fixed-size chunks we need to process 10 pixels: _mm256_mul_ps will potentially read past the end of the allocated buffers If we simply apply the vector instruction twice, we will read past the end of the allocated buffer time-consuming and error-prone, but it is still usually the default choice, since it allows us to work with buffers of any alignment or size. The two algorithms should use exactly the same operations mathematically

ARM? Please remember, that for better performance it is still recommended to make sure that all buffers are aligned to SIMD-word boundary (that is, 32 bytes for AVX and 16 bytes for SSE). Fixed-size chunks we need to process 10 pixels: _mm256_mul_ps will potentially read past the end of the allocated buffers If we simply apply the vector instruction twice, we will read past the end of the allocated buffer time-consuming and error-prone, but it is still usually the default choice, since it allows us to work with buffers of any alignment or size. The two algorithms should use exactly the same operations mathematically

ARM? Please remember, that for better performance it is still recommended to make sure that all buffers are aligned to SIMD-word boundary (that is, 32 bytes for AVX and 16 bytes for SSE). Fixed-size chunks we need to process 10 pixels: _mm256_mul_ps will potentially read past the end of the allocated buffers If we simply apply the vector instruction twice, we will read past the end of the allocated buffer time-consuming and error-prone, but it is still usually the default choice, since it allows us to work with buffers of any alignment or size. The two algorithms should use exactly the same operations mathematically

ARM? Please remember, that for better performance it is still recommended to make sure that all buffers are aligned to SIMD-word boundary (that is, 32 bytes for AVX and 16 bytes for SSE). Fixed-size chunks we need to process 10 pixels: _mm256_mul_ps will potentially read past the end of the allocated buffers If we simply apply the vector instruction twice, we will read past the end of the allocated buffer time-consuming and error-prone, but it is still usually the default choice, since it allows us to work with buffers of any alignment or size. The two algorithms should use exactly the same operations mathematically

ARM? Please remember, that for better performance it is still recommended to make sure that all buffers are aligned to SIMD-word boundary (that is, 32 bytes for AVX and 16 bytes for SSE). Fixed-size chunks we need to process 10 pixels: _mm256_mul_ps will potentially read past the end of the allocated buffers If we simply apply the vector instruction twice, we will read past the end of the allocated buffer time-consuming and error-prone, but it is still usually the default choice, since it allows us to work with buffers of any alignment or size. The two algorithms should use exactly the same operations mathematically

ARM? Please remember, that for better performance it is still recommended to make sure that all buffers are aligned to SIMD-word boundary (that is, 32 bytes for AVX and 16 bytes for SSE). Fixed-size chunks we need to process 10 pixels: _mm256_mul_ps will potentially read past the end of the allocated buffers If we simply apply the vector instruction twice, we will read past the end of the allocated buffer time-consuming and error-prone, but it is still usually the default choice, since it allows us to work with buffers of any alignment or size. The two algorithms should use exactly the same operations mathematically