Added an 'arm64' entry to .travis.yml. (#726)

Details:
- Added a new 'arm64' entry to the .travis.yml file in an attempt to get
  Travis CI to compile both NEON and SVE kernels, even if only NEON
  kernels are exercised in the testing. With this new 'arm64' entry, the
  'cortexa57' entry becomes redundant and may be removed. Thanks to
  RuQing Xu for this suggestion.
- Previously, the macro BLIS_SIMD_MAX_SIZE was *not* being set in
  bli_kernels_arm64.h, which meant that the default value of 64 was
  being used. This caused a runtime consistency check to fail in
  bli_gks.c (in Travis CI), one which requires that

    mr * nr * dt_size > BLIS_STACK_BUF_MAX_SIZE

  for all datatype sizes dt_size, where BLIS_STACK_BUF_MAX_SIZE is
  defined as

    BLIS_SIMD_MAX_NUM_REGISTERS * BLIS_SIMD_MAX_SIZE * 2

  This commit increases BLIS_SIMD_MAX_SIZE to 128 for the 'arm64'
  configuration, thus overriding the default and (hopefully) avoiding
  the aforementioned consistency check failures.
- Appended '|| cat ./output.testsuite' to all 'make' commands in
  travis/do_testsuite.sh. Thanks to RuQing Xu for this suggestion.
- Whitespace changes.

.travis.yml

@@ -62,6 +62,15 @@ matrix:
       CC=aarch64-linux-gnu-gcc-10 CXX=aarch64-linux-gnu-g++-10 \
       PACKAGES="gcc-10-aarch64-linux-gnu g++-10-aarch64-linux-gnu libc6-dev-arm64-cross qemu-system-arm qemu-user" \
       TESTSUITE_WRAPPER="qemu-aarch64 -cpu max,sve=true,sve512=true -L /usr/aarch64-linux-gnu/"
+  # arm64 build and fast testsuite (qemu)
+  # NOTE: This entry omits the -cpu flag so that while both NEON and SVE kernels
+  # are compiled, only NEON kernels will be tested. (h/t to RuQing Xu)
+  - os: linux
+    compiler: aarch64-linux-gnu-gcc-10
+    env: OOT=0 TEST=FAST SDE=0 THR="none" CONF="arm64" \
+      CC=aarch64-linux-gnu-gcc-10 CXX=aarch64-linux-gnu-g++-10 \
+      PACKAGES="gcc-10-aarch64-linux-gnu g++-10-aarch64-linux-gnu libc6-dev-arm64-cross qemu-system-arm qemu-user" \
+      TESTSUITE_WRAPPER="qemu-aarch64 -L /usr/aarch64-linux-gnu/"
 install:
 - if [ "$CC" = "gcc"  ] && [ "$TRAVIS_OS_NAME" = "linux" ]; then export CC="gcc-9"; fi
 - if [ -n "$PACKAGES" ] && [ "$TRAVIS_OS_NAME" = "linux" ]; then sudo apt-get install -y $PACKAGES; fi

config/arm64/bli_family_arm64.h

@@ -39,6 +39,8 @@
 // -- MEMORY ALLOCATION --------------------------------------------------------
 
 #define BLIS_SIMD_ALIGN_SIZE 16
+
+#define BLIS_SIMD_MAX_SIZE 128 // Note: The default is 64.
 #define BLIS_SIMD_MAX_NUM_REGISTERS 32
 
 // SVE-specific configs.

travis/do_testsuite.sh

@@ -9,27 +9,27 @@ export BLIS_JR_NT=1
 export BLIS_IR_NT=1
 
 if [ "$TEST" = "FAST" -o "$TEST" = "ALL" ]; then
-    make testblis-fast
-    $DIST_PATH/testsuite/check-blistest.sh ./output.testsuite
+	make testblis-fast || cat ./output.testsuite
+	$DIST_PATH/testsuite/check-blistest.sh ./output.testsuite
 fi
 
 if [ "$TEST" = "MD" -o "$TEST" = "ALL" ]; then
-	make testblis-md
-    $DIST_PATH/testsuite/check-blistest.sh ./output.testsuite
+	make testblis-md || cat ./output.testsuite
+	$DIST_PATH/testsuite/check-blistest.sh ./output.testsuite
 fi
 
 if [ "$TEST" = "SALT" -o "$TEST" = "ALL" ]; then
 	# Disable multithreading within BLIS.
 	export BLIS_JC_NT=1 BLIS_IC_NT=1 BLIS_JR_NT=1 BLIS_IR_NT=1
-	make testblis-salt
-    $DIST_PATH/testsuite/check-blistest.sh ./output.testsuite
+	make testblis-salt || cat ./output.testsuite
+	$DIST_PATH/testsuite/check-blistest.sh ./output.testsuite
 fi
 
 if [ "$TEST" = "1" -o "$TEST" = "ALL" ]; then
-    make testblis
-    $DIST_PATH/testsuite/check-blistest.sh ./output.testsuite
+	make testblis || cat ./output.testsuite
+	$DIST_PATH/testsuite/check-blistest.sh ./output.testsuite
 fi
 
-make testblas
+make testblas || cat ./output.testsuite
 $DIST_PATH/blastest/check-blastest.sh