Remove unnecessary build step for some workflows (rust-lang#3124)

Workflows that run the perf regressions (`scripts/kani-perf.sh`) do not
need to build Kani, as the perf script itself builds Kani in release
mode:


https://github.com/model-checking/kani/blob/dc0978043c52492112e4ad37a617fd3c8100ef1f/scripts/kani-perf.sh#L12

The extra build step was causing Kani to be built two extra times for
the `benchcomp` flow because the workflow was building it in debug mode,
so the script will end up rebuilding it in release mode. This should
save about 3 minutes for the `benchcomp` flow.

.github/workflows/bench.yml

@@ -55,12 +55,6 @@ jobs:
           os: ubuntu-20.04
           kani_dir: new
 
-      - name: Build Kani (new variant)
-        run: pushd new && cargo build-dev
-
-      - name: Build Kani (old variant)
-        run: pushd old && cargo build-dev
-
       - name: Copy benchmarks from new to old
         run: rm -rf ./old/tests/perf ; cp -r ./new/tests/perf ./old/tests/
 

.github/workflows/cbmc-latest.yml

@@ -32,10 +32,6 @@ jobs:
           os: ${{ matrix.os }}
           kani_dir: 'kani'
 
-      - name: Build Kani
-        working-directory: ./kani
-        run: cargo build-dev
-
       - name: Checkout CBMC under "cbmc"
         uses: actions/checkout@v4
         with:

.github/workflows/kani-m1.yml

@@ -23,8 +23,5 @@ jobs:
         with:
           os: macos-13-xlarge
 
-      - name: Build Kani
-        run: cargo build-dev
-
       - name: Execute Kani regression
         run: ./scripts/kani-regression.sh

.github/workflows/kani.yml

@@ -27,9 +27,6 @@ jobs:
         with:
           os: ${{ matrix.os }}
 
-      - name: Build Kani
-        run: cargo build-dev
-
       - name: Execute Kani regression
         run: ./scripts/kani-regression.sh
 
@@ -88,9 +85,6 @@ jobs:
         with:
           os: ubuntu-20.04
 
-      - name: Build Kani using release mode
-        run: cargo build-dev -- --release
-
       - name: Execute Kani performance tests
         run: ./scripts/kani-perf.sh
         env: