diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml
index 70de682d6..f99192d44 100644
--- a/.github/workflows/release.yml
+++ b/.github/workflows/release.yml
@@ -12,6 +12,8 @@ jobs:
       matrix:
         config:
           - blastoise
+          - machamp
+          - sandslash
     outputs:
       cache-key: ${{ steps.build.outputs.cache-key }}
     steps:
@@ -24,25 +26,37 @@ jobs:
         run: |
           closure="$(nix build -L '.#t1.${{ matrix.config }}.release.docker-layers.final-image' --no-link --print-out-paths)"
           echo "path: $closure"
-          cp "$closure/image.tar" /tmp/t1-image.tar
-          echo "cache-key=$(nix hash file --base32 /tmp/t1-image.tar)" > $GITHUB_OUTPUT
+          cp "$closure/image.tar" /tmp/t1-${{ matrix.config }}-image.tar
+          echo "cache-key-${{ matrix.config }}=$(nix hash file --base32 $closure/image.tar)" > $GITHUB_OUTPUT
+          nix build -L '.#t1.${{ matrix.config }}.release.doc' --out-link docs
       - name: Upload to cache
         uses: actions/cache/save@v4
         with:
-          path: /tmp/t1-image.tar
-          key: ${{ steps.build.outputs.cache-key }}
+          path: /tmp/t1-${{ matrix.config }}-image.tar
+          key: ${{ steps.build.outputs[format('cache-key-{0}', matrix.config)] }}
+      - uses: actions/upload-artifact@v4
+        with:
+          name: docs-${{ matrix.config }}.pdf
+          path: |
+            docs/*.pdf
 
   upload:
     runs-on: ubuntu-latest
     needs: [build]
+    strategy:
+      matrix:
+        config:
+          - blastoise
+          - machamp
+          - sandslash
     steps:
-      - name: Restore from cache
+      - name: Restore ${{ matrix.config }} from cache
         uses: actions/cache/restore@v4
         id: cache
         with:
-          path: /tmp/t1-image.tar
+          path: /tmp/t1-${{ matrix.config }}-image.tar
           fail-on-cache-miss: true
-          key: ${{ needs.build.outputs.cache-key }}
+          key: ${{ steps.build.outputs[format('cache-key-{0}', matrix.config)] }}
       - name: Login to GHCR dot IO
         uses: docker/login-action@v2
         with:
@@ -51,6 +65,6 @@ jobs:
           password: ${{ secrets.GITHUB_TOKEN }}
       - name: Load and push
         run: |
-          docker load < /tmp/t1-image.tar
-          docker tag t1/release:latest ghcr.io/chipsalliance/t1:latest
-          docker push ghcr.io/chipsalliance/t1:latest
+          docker load < /tmp/t1-${{ matrix.config }}-image.tar
+          docker tag chipsalliance/t1-${{ matrix.config }}:latest ghcr.io/chipsalliance/t1-${{ matrix.config }}:latest
+          docker push ghcr.io/chipsalliance/t1-${{ matrix.config }}:latest
diff --git a/nix/t1/default.nix b/nix/t1/default.nix
index 373155df5..e720595ec 100644
--- a/nix/t1/default.nix
+++ b/nix/t1/default.nix
@@ -85,9 +85,7 @@ lib.makeScope newScope
         rtl = innerSelf.callPackage ./rtl.nix { mlirbc = innerSelf.subsystem.mlirbc; };
       };
 
-      release = innerSelf.callPackage ./release {
-        ip-emulator = ip.emu-trace;
-      };
+      release = innerSelf.callPackage ./release { };
     })
   )
   )
diff --git a/nix/t1/release/default.nix b/nix/t1/release/default.nix
index 0f80a8cd5..3b0ebe9cc 100644
--- a/nix/t1/release/default.nix
+++ b/nix/t1/release/default.nix
@@ -10,13 +10,13 @@
 
 , configName
 , t1-script
-, ip-emulator
+, ip
 , cases
 , elaborateConfigJson
 }:
 
 lib.makeScope newScope (scope: rec {
-  inherit ip-emulator elaborateConfigJson;
+  inherit elaborateConfigJson configName;
 
   testCases = with cases; [
     intrinsic.matmul
@@ -34,7 +34,13 @@ lib.makeScope newScope (scope: rec {
       makeWrapper ${t1-script}/bin/t1-helper $out/bin/ip-emulator \
         --add-flags "ipemu" \
         --add-flags "--config ${elaborateConfigJson}" \
-        --add-flags "--emulator-path ${ip-emulator}/bin/emulator"
+        --add-flags "--emulator-path ${ip.emu}/bin/emulator"
+
+      makeWrapper ${t1-script}/bin/t1-helper $out/bin/ip-emulator-trace \
+        --add-flags "ipemu" \
+        --add-flags "--config ${elaborateConfigJson}" \
+        --add-flags "--trace" \
+        --add-flags "--emulator-path ${ip.emu-trace}/bin/emulator"
     '';
 
   docker-layers = scope.callPackage ./docker-layers.nix { };
diff --git a/nix/t1/release/doc.typ b/nix/t1/release/doc.typ
index 1b721c581..ad82fa9f1 100644
--- a/nix/t1/release/doc.typ
+++ b/nix/t1/release/doc.typ
@@ -52,6 +52,7 @@ docker run --name t1 -it -v $PWD:/workspace --rm t1/release:latest /bin/bash
 == What is inside
 
 + IP emulator: `/bin/ip-emulator`
++ IP emulator with trace functionality: `/bin/ip-emulator-trace`
 + Softmax & Linear Normalization & Matmul test cases: `/workspace/cases`
 
 == How to run some workload using IP emulator
@@ -64,5 +65,5 @@ ls /workspace/cases
 ip-emulator --case cases/intrinsic-matmul/bin/intrinsic.matmul.elf
 
 # Get waveform trace file
-ip-emulator --trace --case ...
+ip-emulator-trace --case cases/intrinsic-linear_normalization/bin/intrinsic.linear_normalization.elf
 ```
diff --git a/nix/t1/release/docker-layers.nix b/nix/t1/release/docker-layers.nix
index ce731618b..e8c164f90 100644
--- a/nix/t1/release/docker-layers.nix
+++ b/nix/t1/release/docker-layers.nix
@@ -13,12 +13,13 @@
 , rv32-stdenv
 , emulator-wrapped
 , testCases
+, configName
 }:
 
 let
   # Don't use buildImage which relies on KVM feature
   self = dockerTools.streamLayeredImage {
-    name = "t1/release";
+    name = "chipsalliance/t1-${configName}";
     tag = "latest";
 
     contents = with dockerTools; [
diff --git a/script/src/Main.scala b/script/src/Main.scala
index a8feca15e..1daf89e07 100644
--- a/script/src/Main.scala
+++ b/script/src/Main.scala
@@ -115,20 +115,21 @@ object Main:
       caseName: String
   ): os.Path =
     val pathTail =
-      if (os.exists(os.Path(caseName, os.pwd))) then
+      if os.exists(os.Path(caseName, os.pwd)) || os.exists(os.Path(config, os.pwd)) then
         // It is hard to canoncalize user specify path, so here we use date time instead
-        java.time.LocalDateTime
+        val now = java.time.LocalDateTime
           .now()
           .format(
             java.time.format.DateTimeFormatter.ofPattern("yy-MM-dd-HH-mm-ss")
           )
-      else caseName
+        os.RelPath(now)
+      else os.RelPath(s"$config/$caseName")
 
     val path =
       if (outputDir.isEmpty) then
         if (outputBaseDir.isEmpty) then
-          os.pwd / "testrun" / s"${emuType}emu" / config / pathTail
-        else os.Path(outputBaseDir.get, os.pwd) / config / pathTail
+          os.pwd / "testrun" / s"${emuType}emu" / pathTail
+        else os.Path(outputBaseDir.get, os.pwd) / pathTail
       else os.Path(outputDir.get)
 
     os.makeDir.all(path)
@@ -142,6 +143,7 @@ object Main:
   @main def ipemu(
       @arg(
         name = "case",
+        short = 'C',
         doc = "name alias for loading test case"
       ) testCase: String,
       @arg(
@@ -206,7 +208,8 @@ object Main:
         doc = "Force using x86_64 as cross compiling host platform"
       ) forceX86: Boolean = false,
       @arg(
-        name = "dump-cycle",
+        name = "dump-from-cycle",
+        short = 'D',
         doc = "Specify the dump starting point"
       ) dumpCycle: String = "0.0",
       @arg(