CUDA: use async data loading for FlashAttention (ggml-org#11894)

* CUDA: use async data loading for FlashAttention

---------

Co-authored-by: Diego Devesa <[email protected]>

update release requirements (ggml-org#11897)

server : fix divide-by-zero in metrics reporting (ggml-org#11915)

vulkan: implement several ops relevant for ggml_opt (ggml-org#11769)

* vulkan: support memset_tensor

* vulkan: support GGML_OP_SUM

* vulkan: implement GGML_OP_ARGMAX

* vulkan: implement GGML_OP_SUB

* vulkan: implement GGML_OP_COUNT_EQUAL

* vulkan: implement GGML_OP_OPT_STEP_ADAMW

* vulkan: fix check_results RWKV_WKV6 crash and memory leaks

* vulkan: implement GGML_OP_REPEAT_BACK

* tests: remove invalid test-backend-ops REPEAT_BACK tests

* vulkan: fix COUNT_EQUAL memset using a fillBuffer command

server : bump httplib to 0.19.0 (ggml-org#11908)

common : Fix a typo in help (ggml-org#11899)

This patch fixes a typo in command help.
prefx -> prefix

Signed-off-by: Masanari Iida <[email protected]>

vulkan: support multi/vision rope, and noncontiguous rope (ggml-org#1…

…1902)

metal : fix the crash caused by the lack of residency set support on …

…Intel Macs. (ggml-org#11904)

metal : optimize dequant q6_K kernel (ggml-org#11892)

repo : update links to new url (ggml-org#11886)

* repo : update links to new url

ggml-ci

* cont : more urls

ggml-ci