* [gentoo-commits] repo/proj/guru:dev commit in: app-misc/ollama/
@ 2024-09-19 11:27 Ivan Lloro
0 siblings, 0 replies; 6+ messages in thread
From: Ivan Lloro @ 2024-09-19 11:27 UTC (permalink / raw
To: gentoo-commits
commit: 9be74c9532e61b6673bb539b8e45fc1023c62559
Author: Ivan Lloro <ivan.lloro.boada <AT> gmail <DOT> com>
AuthorDate: Thu Sep 19 11:27:08 2024 +0000
Commit: Ivan Lloro <ivan.lloro.boada <AT> gmail <DOT> com>
CommitDate: Thu Sep 19 11:27:25 2024 +0000
URL: https://gitweb.gentoo.org/repo/proj/guru.git/commit/?id=9be74c95
app-misc/ollama: Add missing keyword for AMD64
Signed-off-by: Ivan Lloro <ivan.lloro.boada <AT> gmail.com>
app-misc/ollama/ollama-9999.ebuild | 1 +
1 file changed, 1 insertion(+)
diff --git a/app-misc/ollama/ollama-9999.ebuild b/app-misc/ollama/ollama-9999.ebuild
index ded584254..e78c85451 100644
--- a/app-misc/ollama/ollama-9999.ebuild
+++ b/app-misc/ollama/ollama-9999.ebuild
@@ -10,6 +10,7 @@ HOMEPAGE="https://ollama.com"
EGIT_REPO_URI="https://github.com/ollama/ollama.git"
LICENSE="MIT"
SLOT="0"
+KEYWORDS="~amd64"
IUSE="nvidia amd"
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [gentoo-commits] repo/proj/guru:dev commit in: app-misc/ollama/
@ 2024-09-19 11:58 Anna Vyalkova
0 siblings, 0 replies; 6+ messages in thread
From: Anna Vyalkova @ 2024-09-19 11:58 UTC (permalink / raw
To: gentoo-commits
commit: 6df145813b847f86037fef1602d03752947e5bf4
Author: Anna (cybertailor) Vyalkova <cyber+gentoo <AT> sysrq <DOT> in>
AuthorDate: Thu Sep 19 11:58:14 2024 +0000
Commit: Anna Vyalkova <cyber+gentoo <AT> sysrq <DOT> in>
CommitDate: Thu Sep 19 11:58:14 2024 +0000
URL: https://gitweb.gentoo.org/repo/proj/guru.git/commit/?id=6df14581
Revert "app-misc/ollama: Add missing keyword for AMD64"
No keywords on live ebuilds.
This reverts commit 9be74c9532e61b6673bb539b8e45fc1023c62559.
Signed-off-by: Anna (cybertailor) Vyalkova <cyber+gentoo <AT> sysrq.in>
app-misc/ollama/ollama-9999.ebuild | 1 -
1 file changed, 1 deletion(-)
diff --git a/app-misc/ollama/ollama-9999.ebuild b/app-misc/ollama/ollama-9999.ebuild
index e78c85451..ded584254 100644
--- a/app-misc/ollama/ollama-9999.ebuild
+++ b/app-misc/ollama/ollama-9999.ebuild
@@ -10,7 +10,6 @@ HOMEPAGE="https://ollama.com"
EGIT_REPO_URI="https://github.com/ollama/ollama.git"
LICENSE="MIT"
SLOT="0"
-KEYWORDS="~amd64"
IUSE="nvidia amd"
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [gentoo-commits] repo/proj/guru:dev commit in: app-misc/ollama/
@ 2024-12-12 19:36 Eric Joldasov
0 siblings, 0 replies; 6+ messages in thread
From: Eric Joldasov @ 2024-12-12 19:36 UTC (permalink / raw
To: gentoo-commits
commit: 2289aca2d83518a11721ead426e4142f7600b6de
Author: Eric Joldasov <bratishkaerik <AT> landless-city <DOT> net>
AuthorDate: Thu Dec 12 19:34:04 2024 +0000
Commit: Eric Joldasov <bratishkaerik <AT> getgoogleoff <DOT> me>
CommitDate: Thu Dec 12 19:34:04 2024 +0000
URL: https://gitweb.gentoo.org/repo/proj/guru.git/commit/?id=2289aca2
app-misc/ollama: bump minimum Go version to 1.23.4
Fixes this error:
```
go mod vendor
go: go.mod requires go >= 1.23.4 (running go 1.23.3; GOTOOLCHAIN=local)
* ERROR: app-misc/ollama-9999::guru failed (unpack phase):
* go mod vendor failed
```
Signed-off-by: Eric Joldasov <bratishkaerik <AT> landless-city.net>
app-misc/ollama/ollama-9999.ebuild | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/app-misc/ollama/ollama-9999.ebuild b/app-misc/ollama/ollama-9999.ebuild
index ded584254..c45ce6165 100644
--- a/app-misc/ollama/ollama-9999.ebuild
+++ b/app-misc/ollama/ollama-9999.ebuild
@@ -19,7 +19,7 @@ RDEPEND="
"
IDEPEND="${RDEPEND}"
BDEPEND="
- >=dev-lang/go-1.21.0
+ >=dev-lang/go-1.23.4
>=dev-build/cmake-3.24
>=sys-devel/gcc-11.4.0
nvidia? ( dev-util/nvidia-cuda-toolkit )
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [gentoo-commits] repo/proj/guru:dev commit in: app-misc/ollama/
@ 2025-01-26 9:04 Yahor Berdnikau
0 siblings, 0 replies; 6+ messages in thread
From: Yahor Berdnikau @ 2025-01-26 9:04 UTC (permalink / raw
To: gentoo-commits
commit: ac646f42de00d0fcada1326f887692fedef16879
Author: Yahor Berdnikau <egorr.berd <AT> gmail <DOT> com>
AuthorDate: Sun Jan 26 09:02:59 2025 +0000
Commit: Yahor Berdnikau <egorr.berd <AT> gmail <DOT> com>
CommitDate: Sun Jan 26 09:02:59 2025 +0000
URL: https://gitweb.gentoo.org/repo/proj/guru.git/commit/?id=ac646f42
app-misc/ollama: Fixes build due to the upstream build system changes
Bug: https://bugs.gentoo.org/946268
Signed-off-by: Yahor Berdnikau <egorr.berd <AT> gmail.com>
app-misc/ollama/metadata.xml | 3 +-
app-misc/ollama/ollama-9999.ebuild | 76 +++++++++++++++++++++++++++++---------
2 files changed, 60 insertions(+), 19 deletions(-)
diff --git a/app-misc/ollama/metadata.xml b/app-misc/ollama/metadata.xml
index 352043225..4ebf01228 100644
--- a/app-misc/ollama/metadata.xml
+++ b/app-misc/ollama/metadata.xml
@@ -7,8 +7,7 @@
<description>A copy from https://github.com/MrPenguin07/ebuilds/tree/master/dev-ml/ollama</description>
</maintainer>
<use>
- <flag name="nvidia">Add support of nvidia</flag>
- <flag name="amd">Add support of amd</flag>
+ <flag name="cuda">Enable NVIDIA CUDA support</flag>
</use>
<upstream>
<remote-id type="github">ollama/ollama</remote-id>
diff --git a/app-misc/ollama/ollama-9999.ebuild b/app-misc/ollama/ollama-9999.ebuild
index c45ce6165..f0d2f0be2 100644
--- a/app-misc/ollama/ollama-9999.ebuild
+++ b/app-misc/ollama/ollama-9999.ebuild
@@ -1,9 +1,10 @@
-# Copyright 2024 Gentoo Authors
+# Copyright 2024-2025 Gentoo Authors
# Distributed under the terms of the GNU General Public License v2
EAPI=8
-inherit git-r3 go-module
+ROCM_VERSION=6.1
+inherit git-r3 go-module rocm
DESCRIPTION="Get up and running with Llama 3, Mistral, Gemma, and other language models."
HOMEPAGE="https://ollama.com"
@@ -11,7 +12,18 @@ EGIT_REPO_URI="https://github.com/ollama/ollama.git"
LICENSE="MIT"
SLOT="0"
-IUSE="nvidia amd"
+IUSE="cuda video_cards_amdgpu
+cpu_flags_x86_avx cpu_flags_x86_avx2
+cpu_flags_x86_avx512f cpu_flags_x86_avx512vbmi cpu_flags_x86_avx512_vnni cpu_flags_x86_avx512_bf16
+"
+
+REQUIRED_USE="
+ cpu_flags_x86_avx2? ( cpu_flags_x86_avx )
+ cpu_flags_x86_avx512f? ( cpu_flags_x86_avx2 )
+ cpu_flags_x86_avx512vbmi? ( cpu_flags_x86_avx512f )
+ cpu_flags_x86_avx512_vnni? ( cpu_flags_x86_avx512f )
+ cpu_flags_x86_avx512_bf16? ( cpu_flags_x86_avx512f )
+"
RDEPEND="
acct-group/ollama
@@ -22,15 +34,14 @@ BDEPEND="
>=dev-lang/go-1.23.4
>=dev-build/cmake-3.24
>=sys-devel/gcc-11.4.0
- nvidia? ( dev-util/nvidia-cuda-toolkit )
- amd? (
- sci-libs/clblast
- dev-libs/rocm-opencl-runtime
+ cuda? ( dev-util/nvidia-cuda-toolkit )
+ video_cards_amdgpu? (
+ =sci-libs/hipBLAS-${ROCM_VERSION}*
)
"
pkg_pretend() {
- if use amd; then
+ if use video_cards_amdgpu || use cuda; then
ewarn "WARNING: AMD & Nvidia support in this ebuild are experimental"
einfo "If you run into issues, especially compiling dev-libs/rocm-opencl-runtime"
einfo "you may try the docker image here https://github.com/ROCm/ROCm-docker"
@@ -44,20 +55,51 @@ src_unpack() {
go-module_live_vendor
}
+src_prepare() {
+ default
+
+ if use video_cards_amdgpu; then
+ # --hip-version gets appended to the compile flags which isn't a known flag.
+ # This causes rocm builds to fail because -Wunused-command-line-argument is turned on.
+ # Use nuclear option to fix this.
+ # Disable -Werror's from go modules.
+ find "${S}" -name ".go" -exec sed -i "s/ -Werror / /g" {} + || die
+ fi
+}
+
src_compile() {
- VERSION=$(
- git describe --tags --first-parent --abbrev=7 --long --dirty --always \
- | sed -e "s/^v//g"
- assert
- )
- export GOFLAGS="'-ldflags=-w -s \"-X=github.com/ollama/ollama/version.Version=${VERSION}\"'"
+ CUSTOM_CPU_FLAGS=""
+ use cpu_flags_x86_avx && CUSTOM_CPU_FLAGS+="avx"
+ use cpu_flags_x86_avx2 && CUSTOM_CPU_FLAGS+=",avx2"
+ use cpu_flags_x86_avx512f && CUSTOM_CPU_FLAGS+=",avx512"
+ use cpu_flags_x86_avx512vbmi && CUSTOM_CPU_FLAGS+=",avx512vbmi"
+ use cpu_flags_x86_avx512_vnni && CUSTOM_CPU_FLAGS+=",avx512vnni"
+ use cpu_flags_x86_avx512_bf16 && CUSTOM_CPU_FLAGS+=",avx512bf16"
- ego generate ./...
- ego build .
+ # Build basic ollama executable with cpu features built in
+ export CUSTOM_CPU_FLAGS
+
+ if use video_cards_amdgpu; then
+ export HIP_ARCHS=$(get_amdgpu_flags)
+ export HIP_PATH="/usr"
+ else
+ export OLLAMA_SKIP_ROCM_GENERATE=1
+ fi
+
+ if ! use cuda; then
+ export OLLAMA_SKIP_CUDA_GENERATE=1
+ fi
+ emake dist
}
src_install() {
- dobin ollama
+ dobin dist/linux-${ARCH}/bin/ollama
+
+ if [[ -d "dist/linux-${ARCH}/lib/ollama" ]] ; then
+ insinto /usr/lib
+ doins -r dist/linux-${ARCH}/lib/ollama
+ fi
+
doinitd "${FILESDIR}"/ollama
}
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [gentoo-commits] repo/proj/guru:dev commit in: app-misc/ollama/
@ 2025-01-31 18:41 Yahor Berdnikau
0 siblings, 0 replies; 6+ messages in thread
From: Yahor Berdnikau @ 2025-01-31 18:41 UTC (permalink / raw
To: gentoo-commits
commit: bfaa93dcace907596bee18483e11f9831f5d6cf9
Author: Yahor Berdnikau <egorr.berd <AT> gmail <DOT> com>
AuthorDate: Fri Jan 31 18:40:07 2025 +0000
Commit: Yahor Berdnikau <egorr.berd <AT> gmail <DOT> com>
CommitDate: Fri Jan 31 18:40:59 2025 +0000
URL: https://gitweb.gentoo.org/repo/proj/guru.git/commit/?id=bfaa93dc
app-misc/ollama: add 0.5.7
Bug: https://bugs.gentoo.org/946268
Signed-off-by: Yahor Berdnikau <egorr.berd <AT> gmail.com>
app-misc/ollama/Manifest | 2 +
app-misc/ollama/ollama-0.5.7.ebuild | 113 ++++++++++++++++++++++++++++++++++++
2 files changed, 115 insertions(+)
diff --git a/app-misc/ollama/Manifest b/app-misc/ollama/Manifest
new file mode 100644
index 000000000..0d00a5bb0
--- /dev/null
+++ b/app-misc/ollama/Manifest
@@ -0,0 +1,2 @@
+DIST ollama-0.5.7-deps.tar.xz 115361652 BLAKE2B 8694b35ef2545481c2e1f6a9d7c462177f25e78aab79288197ec4b33190a905c7e92e54734725cde7f122e1161cbdaf9c72ae40cbeb0b8ee0af9075e3dbb6691 SHA512 d7abe6266120cb9e731e53f6e14e16d8991e76fbab06348de945f53dc407788324b850308e7200616d92dd17a417ad0a2491eddd543bbe2cfee6a66c8ab81840
+DIST ollama-0.5.7.gh.tar.gz 2345089 BLAKE2B 7d5063e9f665ab2a957d449b38017e6a9bb435c938749161c711cfc35d8a0361e7f4db214e0782f3b51c70c909fc5be8b76ca342cda6163b5aca5fdd733c55d9 SHA512 ea8adcec4f8f932c422a400b8cafb4b983bfa0721cd14383ceb8e0a4f588ecd1289d2e1de46a916c1b34d13e5dab2825ef11a37fc3e797345348dea3bd9144fe
diff --git a/app-misc/ollama/ollama-0.5.7.ebuild b/app-misc/ollama/ollama-0.5.7.ebuild
new file mode 100644
index 000000000..408c054e1
--- /dev/null
+++ b/app-misc/ollama/ollama-0.5.7.ebuild
@@ -0,0 +1,113 @@
+# Copyright 2024-2025 Gentoo Authors
+# Distributed under the terms of the GNU General Public License v2
+
+EAPI=8
+
+ROCM_VERSION=6.1
+inherit go-module rocm
+
+DESCRIPTION="Get up and running with Llama 3, Mistral, Gemma, and other language models."
+HOMEPAGE="https://ollama.com"
+SRC_URI="https://github.com/ollama/${PN}/archive/refs/tags/v${PV}.tar.gz -> ${P}.gh.tar.gz"
+SRC_URI+=" https://github.com/Tapchicoma/ebuild-deps/raw/refs/heads/main/go-deps/${PN}-${PV}-deps.tar.xz"
+S="${WORKDIR}/${PN}-${PV}"
+LICENSE="MIT"
+SLOT="0"
+
+IUSE="cuda video_cards_amdgpu
+cpu_flags_x86_avx cpu_flags_x86_avx2
+cpu_flags_x86_avx512f cpu_flags_x86_avx512vbmi cpu_flags_x86_avx512_vnni cpu_flags_x86_avx512_bf16
+"
+
+REQUIRED_USE="
+ cpu_flags_x86_avx2? ( cpu_flags_x86_avx )
+ cpu_flags_x86_avx512f? ( cpu_flags_x86_avx2 )
+ cpu_flags_x86_avx512vbmi? ( cpu_flags_x86_avx512f )
+ cpu_flags_x86_avx512_vnni? ( cpu_flags_x86_avx512f )
+ cpu_flags_x86_avx512_bf16? ( cpu_flags_x86_avx512f )
+"
+
+RDEPEND="
+ acct-group/ollama
+ acct-user/ollama
+"
+
+DEPEND="
+ >=dev-lang/go-1.23.4
+ >=dev-build/cmake-3.24
+ >=sys-devel/gcc-11.4.0
+ cuda? ( dev-util/nvidia-cuda-toolkit )
+ video_cards_amdgpu? (
+ sci-libs/hipBLAS[${ROCM_USEDEP}]
+ )
+"
+
+pkg_pretend() {
+ if use video_cards_amdgpu; then
+ ewarn "WARNING: AMD support in this ebuild are experimental"
+ einfo "If you run into issues, especially compiling dev-libs/rocm-opencl-runtime"
+ einfo "you may try the docker image here https://github.com/ROCm/ROCm-docker"
+ einfo "and follow instructions here"
+ einfo "https://rocm.docs.amd.com/projects/install-on-linux/en/latest/how-to/docker.html"
+ fi
+}
+
+src_prepare() {
+ default
+
+ if use video_cards_amdgpu; then
+ # --hip-version gets appended to the compile flags which isn't a known flag.
+ # This causes rocm builds to fail because -Wunused-command-line-argument is turned on.
+ # Use nuclear option to fix this.
+ # Disable -Werror's from go modules.
+ find "${S}" -name ".go" -exec sed -i "s/ -Werror / /g" {} + || die
+ fi
+}
+
+src_compile() {
+ CUSTOM_CPU_FLAGS=""
+ use cpu_flags_x86_avx && CUSTOM_CPU_FLAGS+="avx"
+ use cpu_flags_x86_avx2 && CUSTOM_CPU_FLAGS+=",avx2"
+ use cpu_flags_x86_avx512f && CUSTOM_CPU_FLAGS+=",avx512"
+ use cpu_flags_x86_avx512vbmi && CUSTOM_CPU_FLAGS+=",avx512vbmi"
+ use cpu_flags_x86_avx512_vnni && CUSTOM_CPU_FLAGS+=",avx512vnni"
+ use cpu_flags_x86_avx512_bf16 && CUSTOM_CPU_FLAGS+=",avx512bf16"
+
+ # Build basic ollama executable with cpu features built in
+ export CUSTOM_CPU_FLAGS
+
+ if use video_cards_amdgpu; then
+ export HIP_ARCHS=$(get_amdgpu_flags)
+ export HIP_PATH="/usr"
+ else
+ export OLLAMA_SKIP_ROCM_GENERATE=1
+ fi
+
+ if ! use cuda; then
+ export OLLAMA_SKIP_CUDA_GENERATE=1
+ fi
+ emake dist
+}
+
+src_install() {
+ dobin dist/linux-${ARCH}/bin/ollama
+
+ if [[ -d "dist/linux-${ARCH}/lib/ollama" ]] ; then
+ insinto /usr/lib
+ doins -r dist/linux-${ARCH}/lib/ollama
+ fi
+
+ doinitd "${FILESDIR}"/ollama
+}
+
+pkg_preinst() {
+ keepdir /var/log/ollama
+ fowners ollama:ollama /var/log/ollama
+}
+
+pkg_postinst() {
+ einfo "Quick guide:"
+ einfo "ollama serve"
+ einfo "ollama run llama3:70b"
+ einfo "See available models at https://ollama.com/library"
+}
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [gentoo-commits] repo/proj/guru:dev commit in: app-misc/ollama/
@ 2025-01-31 19:13 Yahor Berdnikau
0 siblings, 0 replies; 6+ messages in thread
From: Yahor Berdnikau @ 2025-01-31 19:13 UTC (permalink / raw
To: gentoo-commits
commit: 1879435c4a4365e815a275c8c40069c7b5f99b9a
Author: Yahor Berdnikau <egorr.berd <AT> gmail <DOT> com>
AuthorDate: Fri Jan 31 19:13:09 2025 +0000
Commit: Yahor Berdnikau <egorr.berd <AT> gmail <DOT> com>
CommitDate: Fri Jan 31 19:13:09 2025 +0000
URL: https://gitweb.gentoo.org/repo/proj/guru.git/commit/?id=1879435c
app-misc/ollama: keyword 0.5.7 for ~amd64
Bug: https://bugs.gentoo.org/946268
Signed-off-by: Yahor Berdnikau <egorr.berd <AT> gmail.com>
app-misc/ollama/ollama-0.5.7.ebuild | 2 ++
1 file changed, 2 insertions(+)
diff --git a/app-misc/ollama/ollama-0.5.7.ebuild b/app-misc/ollama/ollama-0.5.7.ebuild
index 408c054e1..f1373d6fa 100644
--- a/app-misc/ollama/ollama-0.5.7.ebuild
+++ b/app-misc/ollama/ollama-0.5.7.ebuild
@@ -14,6 +14,8 @@ S="${WORKDIR}/${PN}-${PV}"
LICENSE="MIT"
SLOT="0"
+KEYWORDS="~amd64"
+
IUSE="cuda video_cards_amdgpu
cpu_flags_x86_avx cpu_flags_x86_avx2
cpu_flags_x86_avx512f cpu_flags_x86_avx512vbmi cpu_flags_x86_avx512_vnni cpu_flags_x86_avx512_bf16
^ permalink raw reply related [flat|nested] 6+ messages in thread
end of thread, other threads:[~2025-01-31 19:13 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-01-26 9:04 [gentoo-commits] repo/proj/guru:dev commit in: app-misc/ollama/ Yahor Berdnikau
-- strict thread matches above, loose matches on Subject: below --
2025-01-31 19:13 Yahor Berdnikau
2025-01-31 18:41 Yahor Berdnikau
2024-12-12 19:36 Eric Joldasov
2024-09-19 11:58 Anna Vyalkova
2024-09-19 11:27 Ivan Lloro
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox