diff --git a/.github/workflows/sycl-linux-precommit.yml b/.github/workflows/sycl-linux-precommit.yml
index 7740117e00634..c70953457bada 100644
--- a/.github/workflows/sycl-linux-precommit.yml
+++ b/.github/workflows/sycl-linux-precommit.yml
@@ -121,7 +121,7 @@ jobs:
       image: ${{ matrix.image }}
       image_options: ${{ matrix.image_options }}
       target_devices: ${{ matrix.target_devices }}
-      extra_lit_opts: --param fallback-to-build-if-requires-build-and-run=True ${{ matrix.extra_lit_opts }}
+      extra_lit_opts: ${{ matrix.extra_lit_opts }}
       reset_intel_gpu: ${{ matrix.reset_intel_gpu }}
       ref: ${{ github.sha }}
       sycl_toolchain_artifact: sycl_linux_default
diff --git a/sycl/test-e2e/AOT/cpu.cpp b/sycl/test-e2e/AOT/cpu.cpp
index e7bedc04b07ba..4d1a24f649e30 100644
--- a/sycl/test-e2e/AOT/cpu.cpp
+++ b/sycl/test-e2e/AOT/cpu.cpp
@@ -7,9 +7,9 @@
 //===---------------------------------------------------------------------===//
 
 // REQUIRES: opencl-aot, cpu
-// REQUIRES: build-and-run-mode
 
-// RUN: %clangxx -fsycl -fsycl-targets=spir64_x86_64 %S/Inputs/aot.cpp -o %t.out
+// CPU AOT targets host isa, so we compile on the run system instead.
+// RUN: %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 %S/Inputs/aot.cpp -o %t.out
 // RUN: %{run} %t.out
 
 // Test that opencl-aot can handle multiple build options.
diff --git a/sycl/test-e2e/AOT/double.cpp b/sycl/test-e2e/AOT/double.cpp
index 4d48f5c7b57ec..0cba8616dd261 100644
--- a/sycl/test-e2e/AOT/double.cpp
+++ b/sycl/test-e2e/AOT/double.cpp
@@ -2,9 +2,10 @@
 // using fp64 can be compiled AOT.
 
 // REQUIRES: ocloc, opencl-aot, any-device-is-cpu
-// REQUIRES: build-and-run-mode
 // RUN: %clangxx -fsycl -fsycl-targets=intel_gpu_tgllp -o %t.tgllp.out %s
-// RUN: %clangxx -fsycl -fsycl-targets=spir64_x86_64 -o %t.x86.out %s
+
+// CPU AOT targets host isa, so we compile on the run system instead.
+// RUN: %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 -o %t.x86.out %s
 // RUN: %if cpu %{ %{run} %t.x86.out %}
 
 // ocloc on windows does not have support for PVC, so this command will
diff --git a/sycl/test-e2e/AOT/half.cpp b/sycl/test-e2e/AOT/half.cpp
index 5711b21735f72..3e6ba389ac69c 100644
--- a/sycl/test-e2e/AOT/half.cpp
+++ b/sycl/test-e2e/AOT/half.cpp
@@ -2,9 +2,10 @@
 // using fp16 can be compiled AOT.
 
 // REQUIRES: ocloc, opencl-aot, any-device-is-cpu
-// REQUIRES: build-and-run-mode
 // RUN: %clangxx -fsycl -fsycl-targets=intel_gpu_tgllp -o %t.tgllp.out %s
-// RUN: %clangxx -fsycl -fsycl-targets=spir64_x86_64 -o %t.x86.out %s
+
+// CPU AOT targets host isa, so we compile on the run system instead.
+// RUN: %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 -o %t.x86.out %s
 // RUN: %if cpu %{ %{run} %t.x86.out %}
 
 // ocloc on windows does not have support for PVC, so this command will
diff --git a/sycl/test-e2e/Adapters/sycl-targets-order.cpp b/sycl/test-e2e/Adapters/sycl-targets-order.cpp
index 273d3f51ec328..1690e1f247daa 100644
--- a/sycl/test-e2e/Adapters/sycl-targets-order.cpp
+++ b/sycl/test-e2e/Adapters/sycl-targets-order.cpp
@@ -5,8 +5,7 @@
 // RUN: %{run-unfiltered-devices} env ONEAPI_DEVICE_SELECTOR="opencl:*" %t-nvptx64-spir64.out
 // RUN: %{run-unfiltered-devices} env ONEAPI_DEVICE_SELECTOR="cuda:*"   %t-nvptx64-spir64.out
 
-// REQUIRES: opencl, cuda
-// REQUIRES: build-and-run-mode
+// REQUIRES: opencl, target-spir, any-triple-is-nvidia
 
 //==------- sycl-targets-order.cpp - SYCL -fsycl-targets order test --------==//
 //
@@ -36,7 +35,7 @@ int main(int argc, char **argv) {
   sycl::buffer<unsigned int, 1> buffer(4);
 
   // size of the index space for the kernel
-  sycl::range<1> NumOfWorkItems{buffer.get_count()};
+  sycl::range<1> NumOfWorkItems{buffer.size()};
 
   // submit a command group(work) to the queue
   queue.submit([&](sycl::handler &cgh) {
@@ -55,7 +54,7 @@ int main(int argc, char **argv) {
 
   // check the results
   bool mismatch = false;
-  for (unsigned int i = 0; i < buffer.get_count(); ++i) {
+  for (unsigned int i = 0; i < buffer.size(); ++i) {
     if (host_accessor[i] != i) {
       std::cout << "The result is incorrect for element: " << i
                 << " , expected: " << i << " , got: " << host_accessor[i]
diff --git a/sycl/test-e2e/AddressSanitizer/aot/cpu.cpp b/sycl/test-e2e/AddressSanitizer/aot/cpu.cpp
index e528574bcf1f6..41a18da31acaf 100644
--- a/sycl/test-e2e/AddressSanitizer/aot/cpu.cpp
+++ b/sycl/test-e2e/AddressSanitizer/aot/cpu.cpp
@@ -1,14 +1,13 @@
 // REQUIRES: linux, opencl-aot, cpu
-// REQUIRES: build-and-run-mode
 
-// RUN: %{build} %device_asan_aot_flags -O0 -g %S/Inputs/host-usm-oob.cpp -o %t.out
+// RUN: %{run-aux} %{build} %device_asan_aot_flags -O0 -g %S/Inputs/host-usm-oob.cpp -o %t.out
 // RUN: %{run} not %t.out 2>&1 | FileCheck %S/Inputs/host-usm-oob.cpp
 
-// RUN: %{build} %device_asan_aot_flags -O1 -g %S/Inputs/host-usm-oob.cpp -o %t.out
+// RUN: %{run-aux} %{build} %device_asan_aot_flags -O1 -g %S/Inputs/host-usm-oob.cpp -o %t.out
 // RUN: %{run} not %t.out 2>&1 | FileCheck %S/Inputs/host-usm-oob.cpp
 
-// RUN: %{build} %device_asan_aot_flags -O2 -g %S/Inputs/host-usm-oob.cpp -o %t.out
+// RUN: %{run-aux} %{build} %device_asan_aot_flags -O2 -g %S/Inputs/host-usm-oob.cpp -o %t.out
 // RUN: %{run} not %t.out 2>&1 | FileCheck %S/Inputs/host-usm-oob.cpp
 
-// RUN: %{build} %device_asan_aot_flags -O3 -g %S/Inputs/host-usm-oob.cpp -o %t.out
+// RUN: %{run-aux} %{build} %device_asan_aot_flags -O3 -g %S/Inputs/host-usm-oob.cpp -o %t.out
 // RUN: %{run} not %t.out 2>&1 | FileCheck %S/Inputs/host-usm-oob.cpp
diff --git a/sycl/test-e2e/AddressSanitizer/aot/gpu.cpp b/sycl/test-e2e/AddressSanitizer/aot/gpu.cpp
index 648fe4b0d41d7..c45701af54543 100644
--- a/sycl/test-e2e/AddressSanitizer/aot/gpu.cpp
+++ b/sycl/test-e2e/AddressSanitizer/aot/gpu.cpp
@@ -1,5 +1,4 @@
 // REQUIRES: linux, ocloc, gpu && level_zero
-// REQUIRES: build-and-run-mode
 
 // RUN: %{build} %device_asan_aot_flags -O0 -g %S/Inputs/host-usm-oob.cpp -o %t.out
 // RUN: %{run} not %t.out 2>&1 | FileCheck %S/Inputs/host-usm-oob.cpp
diff --git a/sycl/test-e2e/AddressSanitizer/out-of-bounds/DeviceGlobal/device_global.cpp b/sycl/test-e2e/AddressSanitizer/out-of-bounds/DeviceGlobal/device_global.cpp
index eba4f7cfb010d..4145e3741a242 100644
--- a/sycl/test-e2e/AddressSanitizer/out-of-bounds/DeviceGlobal/device_global.cpp
+++ b/sycl/test-e2e/AddressSanitizer/out-of-bounds/DeviceGlobal/device_global.cpp
@@ -15,7 +15,7 @@
 // a choice based on that. This is rather fragile but workarounds the issue for
 // the time being.
 
-// DEFINE: %{not} = not %if test-mode-run-only %{ --crash %}
+// DEFINE: %{not} = not %if run-mode && !build-mode %{ --crash %}
 
 // RUN: %{build} %device_asan_flags -O0 -g -o %t1.out
 // RUN: %{run} %{not} %t1.out 2>&1 | FileCheck %s
diff --git a/sycl/test-e2e/BFloat16/bfloat16_example_aot.cpp b/sycl/test-e2e/BFloat16/bfloat16_example_aot.cpp
index 32c6c3aae8f06..c79198417d6cf 100644
--- a/sycl/test-e2e/BFloat16/bfloat16_example_aot.cpp
+++ b/sycl/test-e2e/BFloat16/bfloat16_example_aot.cpp
@@ -4,7 +4,6 @@
 ///
 
 // REQUIRES: opencl-aot, ocloc, gpu-intel-gen12, any-device-is-cpu
-// REQUIRES: build-and-run-mode
 
 // RUN: %clangxx -fsycl -fsycl-targets=spir64 %s -o %t.out
 // RUN: %{run} %t.out
@@ -12,7 +11,8 @@
 // RUN: %clangxx -fsycl -fsycl-targets=spir64,spir64_gen -Xsycl-target-backend=spir64_gen "-device gen12lp" %s -o %t.out
 // RUN: %{run} %t.out
 
-// RUN: %clangxx -fsycl -fsycl-targets=spir64_x86_64,spir64_gen -Xsycl-target-backend=spir64_gen "-device gen12lp" %s -o %t.out
+// CPU AOT targets host isa, so we compile on the run system instead.
+// RUN: %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64,spir64_gen -Xsycl-target-backend=spir64_gen "-device gen12lp" %s -o %t.out
 // RUN: %{run} %t.out
 
 #include "bfloat16_example.hpp"
diff --git a/sycl/test-e2e/BFloat16/bfloat16_example_aot_cpu.cpp b/sycl/test-e2e/BFloat16/bfloat16_example_aot_cpu.cpp
index 9231a62428272..0237636b51c71 100644
--- a/sycl/test-e2e/BFloat16/bfloat16_example_aot_cpu.cpp
+++ b/sycl/test-e2e/BFloat16/bfloat16_example_aot_cpu.cpp
@@ -4,12 +4,12 @@
 ///
 
 // REQUIRES: opencl-aot, ocloc, gpu-intel-gen12, any-device-is-cpu
-// REQUIRES: build-and-run-mode
 
 // RUN: %clangxx -fsycl -fsycl-targets=spir64,spir64_gen -Xsycl-target-backend=spir64_gen "-device dg1" %s -o %t.out
 // RUN: %if cpu %{ %{run} %t.out %}
 
-// RUN: %clangxx -fsycl -fsycl-targets=spir64_x86_64,spir64_gen -Xsycl-target-backend=spir64_gen "-device dg1" %s -o %t.out
+// CPU AOT targets host isa, so we compile on the run system instead.
+// RUN: %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64,spir64_gen -Xsycl-target-backend=spir64_gen "-device dg1" %s -o %t.out
 // RUN: %if cpu %{ %{run} %t.out %}
 
 #include "bfloat16_example.hpp"
diff --git a/sycl/test-e2e/Compression/compression_aot.cpp b/sycl/test-e2e/Compression/compression_aot.cpp
index ba95974fde781..115458559678e 100644
--- a/sycl/test-e2e/Compression/compression_aot.cpp
+++ b/sycl/test-e2e/Compression/compression_aot.cpp
@@ -1,6 +1,6 @@
 // End-to-End test for testing device image compression in AOT.
 // REQUIRES: zstd, opencl-aot, cpu
-// REQUIRES: build-and-run-mode
 
-// RUN: %clangxx -fsycl -fsycl-targets=spir64_x86_64 %O0 --offload-compress --offload-compression-level=3 %S/Inputs/single_kernel.cpp -o %t_compress.out
+// CPU AOT targets host isa, so we compile on the run system instead.
+// RUN: %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 %O0 --offload-compress --offload-compression-level=3 %S/Inputs/single_kernel.cpp -o %t_compress.out
 // RUN: %{run} %t_compress.out
diff --git a/sycl/test-e2e/Compression/compression_separate_compile.cpp b/sycl/test-e2e/Compression/compression_separate_compile.cpp
index afbe442230e29..db268153d21e2 100644
--- a/sycl/test-e2e/Compression/compression_separate_compile.cpp
+++ b/sycl/test-e2e/Compression/compression_separate_compile.cpp
@@ -2,27 +2,27 @@
 // seperatly compile and link device images.
 
 // REQUIRES: zstd, opencl-aot, cpu, linux
-// REQUIRES: build-and-run-mode
 
+// CPU AOT targets host isa, so we compile everything on the run system instead.
 //////////////////////  Compile device images
-// RUN: %clangxx -fsycl -fsycl-targets=spir64_x86_64 -fsycl-host-compiler=clang++ -fsycl-host-compiler-options='-std=c++17 -Wno-attributes -Wno-deprecated-declarations -fPIC -DENABLE_KERNEL1' -DENABLE_KERNEL1 -c %s -o %t_kernel1_aot.o
-// RUN: %clangxx -fsycl -fsycl-targets=spir64_x86_64 -fsycl-host-compiler=clang++ -fsycl-host-compiler-options='-std=c++17 -Wno-attributes -Wno-deprecated-declarations -fPIC -DENABLE_KERNEL2' -DENABLE_KERNEL2 -c %s -o %t_kernel2_aot.o
+// RUN: %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 -fsycl-host-compiler=clang++ -fsycl-host-compiler-options='-std=c++17 -Wno-attributes -Wno-deprecated-declarations -fPIC -DENABLE_KERNEL1' -DENABLE_KERNEL1 -c %s -o %t_kernel1_aot.o
+// RUN: %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 -fsycl-host-compiler=clang++ -fsycl-host-compiler-options='-std=c++17 -Wno-attributes -Wno-deprecated-declarations -fPIC -DENABLE_KERNEL2' -DENABLE_KERNEL2 -c %s -o %t_kernel2_aot.o
 
 //////////////////////   Link device images
-// RUN: %clangxx --offload-compress -fsycl -fsycl-link -fsycl-targets=spir64_x86_64 -fPIC %t_kernel1_aot.o %t_kernel2_aot.o -o %t_compressed_image.o -v
+// RUN: %{run-aux} %clangxx --offload-compress -fsycl -fsycl-link -fsycl-targets=spir64_x86_64 -fPIC %t_kernel1_aot.o %t_kernel2_aot.o -o %t_compressed_image.o -v
 
 // Make sure the clang-offload-wrapper is called with the --offload-compress
 // option.
-// RUN: %clangxx --offload-compress -fsycl -fsycl-link -fsycl-targets=spir64_x86_64 -fPIC %t_kernel1_aot.o %t_kernel2_aot.o -o %t_compressed_image.o -### &> %t_driver_opts.txt
-// RUN: FileCheck -input-file=%t_driver_opts.txt %s --check-prefix=CHECK-DRIVER-OPTS
+// RUN: %{run-aux} %clangxx --offload-compress -fsycl -fsycl-link -fsycl-targets=spir64_x86_64 -fPIC %t_kernel1_aot.o %t_kernel2_aot.o -o %t_compressed_image.o -### &> %t_driver_opts.txt
+// RUN: %{run-aux} FileCheck -input-file=%t_driver_opts.txt %s --check-prefix=CHECK-DRIVER-OPTS
 
 // CHECK-DRIVER-OPTS: clang-offload-wrapper{{.*}} "-offload-compress"
 
 //////////////////////   Compile the host program
-// RUN: %clangxx -fsycl -std=c++17 -Wno-attributes -Wno-deprecated-declarations -fPIC -c %s -o %t_main.o
+// RUN: %{run-aux} %clangxx -fsycl -std=c++17 -Wno-attributes -Wno-deprecated-declarations -fPIC -c %s -o %t_main.o
 
 //////////////////////   Link the host program and compressed device images
-// RUN: %clangxx -fsycl %t_main.o %t_kernel1_aot.o %t_kernel2_aot.o %t_compressed_image.o -o %t_compress.out
+// RUN: %{run-aux} %clangxx -fsycl %t_main.o %t_kernel1_aot.o %t_kernel2_aot.o %t_compressed_image.o -o %t_compress.out
 
 // RUN: %{run} %t_compress.out
 
diff --git a/sycl/test-e2e/DeviceArchitecture/device_architecture_on_device_aot.cpp b/sycl/test-e2e/DeviceArchitecture/device_architecture_on_device_aot.cpp
index 1d4c41f117136..179b11cdd3bca 100644
--- a/sycl/test-e2e/DeviceArchitecture/device_architecture_on_device_aot.cpp
+++ b/sycl/test-e2e/DeviceArchitecture/device_architecture_on_device_aot.cpp
@@ -1,6 +1,7 @@
 // REQUIRES: opencl-aot, cpu
-// REQUIRES: build-and-run-mode
-// RUN: %clangxx -fsycl -fsycl-targets=spir64_x86_64 %s -o %t.out
+
+// CPU AOT targets host isa, so we compile on the run system instead.
+// RUN: %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 %s -o %t.out
 // RUN: %{run} %t.out
 
 #include <sycl/detail/core.hpp>
diff --git a/sycl/test-e2e/DeviceCodeSplit/aot-cpu.cpp b/sycl/test-e2e/DeviceCodeSplit/aot-cpu.cpp
index 319c69af0faef..2b0907501efdc 100644
--- a/sycl/test-e2e/DeviceCodeSplit/aot-cpu.cpp
+++ b/sycl/test-e2e/DeviceCodeSplit/aot-cpu.cpp
@@ -1,6 +1,6 @@
 // REQUIRES: opencl-aot, cpu
-// REQUIRES: build-and-run-mode
 
-// RUN: %clangxx -fsycl -fsycl-device-code-split=per_source -fsycl-targets=spir64_x86_64 -I %S/Inputs -o %t.out %S/split-per-source-main.cpp %S/Inputs/split-per-source-second-file.cpp \
-// RUN: -fsycl-dead-args-optimization
+// CPU AOT targets host isa, so we compile on the run system instead.
+// RUN: %{run-aux} %clangxx -fsycl -fsycl-device-code-split=per_source -fsycl-targets=spir64_x86_64 -I %S/Inputs -o %t.out \
+// RUN:   %S/split-per-source-main.cpp %S/Inputs/split-per-source-second-file.cpp -fsycl-dead-args-optimization
 // RUN: %{run} %t.out
diff --git a/sycl/test-e2e/DeviceLib/assert-aot.cpp b/sycl/test-e2e/DeviceLib/assert-aot.cpp
index f5c8c61f07d78..adee4e8e124f6 100644
--- a/sycl/test-e2e/DeviceLib/assert-aot.cpp
+++ b/sycl/test-e2e/DeviceLib/assert-aot.cpp
@@ -1,5 +1,5 @@
 // REQUIRES: opencl-aot, cpu, linux
-// REQUIRES: build-and-run-mode
 
-// RUN: %clangxx -DSYCL_FALLBACK_ASSERT=1 -fsycl -fsycl-targets=spir64_x86_64 %S/assert.cpp -o %t.aot.out
+// CPU AOT targets host isa, so we compile on the run system instead.
+// RUN: %{run-aux} %clangxx -DSYCL_FALLBACK_ASSERT=1 -fsycl -fsycl-targets=spir64_x86_64 %S/assert.cpp -o %t.aot.out
 // RUN: env EXPECTED_SIGNAL=SIGABRT SHOULD_CRASH=1 %{run} %t.aot.out 2>&1 | FileCheck %S/assert.cpp --check-prefixes=CHECK-MESSAGE
diff --git a/sycl/test-e2e/DeviceLib/cmath-aot.cpp b/sycl/test-e2e/DeviceLib/cmath-aot.cpp
index c9897f652f034..6c1eed40ce4ec 100644
--- a/sycl/test-e2e/DeviceLib/cmath-aot.cpp
+++ b/sycl/test-e2e/DeviceLib/cmath-aot.cpp
@@ -1,17 +1,17 @@
 // REQUIRES: opencl-aot, cpu
-// REQUIRES: build-and-run-mode
 // UNSUPPORTED: windows
 
 // DEFINE: %{mathflags} = %if cl_options %{/clang:-fno-fast-math%} %else %{-fno-fast-math%}
 
-// RUN: %clangxx -fsycl -fsycl-targets=spir64_x86_64 %{mathflags} %S/cmath_test.cpp -o %t.cmath.out
+// CPU AOT targets host isa, so we compile on the run system instead.
+// RUN: %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 %{mathflags} %S/cmath_test.cpp -o %t.cmath.out
 // RUN: %{run} %t.cmath.out
 
-// RUN: %clangxx -fsycl -fsycl-targets=spir64_x86_64 %{mathflags} %S/cmath_fp64_test.cpp -o %t.cmath.fp64.out
+// RUN: %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 %{mathflags} %S/cmath_fp64_test.cpp -o %t.cmath.fp64.out
 // RUN: %{run} %t.cmath.fp64.out
 
-// RUN: %clangxx -fsycl -fsycl-targets=spir64_x86_64 %{mathflags} %S/std_complex_math_test.cpp -o %t.complex.out
+// RUN: %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 %{mathflags} %S/std_complex_math_test.cpp -o %t.complex.out
 // RUN: %{run} %t.complex.out
 
-// RUN: %clangxx -fsycl -fsycl-targets=spir64_x86_64 %{mathflags} %S/std_complex_math_fp64_test.cpp -o %t.complex.fp64.out
+// RUN: %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 %{mathflags} %S/std_complex_math_fp64_test.cpp -o %t.complex.fp64.out
 // RUN: %{run} %t.complex.fp64.out
diff --git a/sycl/test-e2e/E2EExpr.py b/sycl/test-e2e/E2EExpr.py
index 00856bf0bb1f5..1b0256f393b27 100644
--- a/sycl/test-e2e/E2EExpr.py
+++ b/sycl/test-e2e/E2EExpr.py
@@ -3,7 +3,6 @@
 
 class E2EExpr(BooleanExpression):
     build_specific_features = {
-        "build-and-run-mode",
         "preview-mode",
         "target-spir",
         "target-nvidia",
diff --git a/sycl/test-e2e/KernelAndProgram/test_cache_jit_aot.cpp b/sycl/test-e2e/KernelAndProgram/test_cache_jit_aot.cpp
index 09399274e0472..5a03a70c92e04 100644
--- a/sycl/test-e2e/KernelAndProgram/test_cache_jit_aot.cpp
+++ b/sycl/test-e2e/KernelAndProgram/test_cache_jit_aot.cpp
@@ -2,37 +2,36 @@
 // cannot do that reliably when number of devices is unknown.
 //
 // REQUIRES: level_zero, ocloc
-// REQUIRES: build-and-run-mode
 //
 // DEFINE: %{cache_vars} = env SYCL_CACHE_PERSISTENT=1 SYCL_CACHE_TRACE=1 SYCL_CACHE_DIR=%t/cache_dir
 // DEFINE: %{build_cmd} = %{build}
-// RUN: mkdir -p %t/cache_dir
+// RUN: %{run-aux} mkdir -p %t/cache_dir
 //
 // The following block of code should be copy-pasted as-is to verify different
 // JIT/AOT options. Don't know how to avoid code duplication.
 // ******************************
 // Check the logs first.
-// RUN: %{build_cmd} -DVALUE=1 -o %t.out
-// RUN: rm -rf %t/cache_dir/*
+// RUN: %{run-aux} %{build_cmd} -DVALUE=1 -o %t.out
+// RUN: %{run-aux} rm -rf %t/cache_dir/*
 // RUN: %{cache_vars} %{run-unfiltered-devices} %t.out 2>&1 | FileCheck %s  --check-prefixes=CHECK-CACHE-WRITE
 // RUN: %{cache_vars} %{run-unfiltered-devices} %t.out 2>&1 | FileCheck %s  --check-prefixes=CHECK-CACHE-READ
 // RUN: %{cache_vars} %{run-unfiltered-devices} %t.out 2>&1 | FileCheck %s  --check-prefixes=CHECK-CACHE-READ
 //
 // Now try to substitute the cached image and verify it is actually taken and
 // the code/binary there is executed.
-// RUN: mv %t/cache_dir/*/*/*/*/*.bin %t.value1.bin
-// RUN: rm -rf %t/cache_dir/*
-// RUN: %{build_cmd} -DVALUE=2 -o %t.out
+// RUN: %{run-aux} mv %t/cache_dir/*/*/*/*/*.bin %t.value1.bin
+// RUN: %{run-aux} rm -rf %t/cache_dir/*
+// RUN: %{run-aux} %{build_cmd} -DVALUE=2 -o %t.out
 // RUN: %{cache_vars} %{run-unfiltered-devices} %t.out 2>&1 | FileCheck %s --check-prefixes RESULT2
-// RUN: mv %t.value1.bin %t/cache_dir/*/*/*/*/*.bin
+// RUN: %{run-aux} mv %t.value1.bin %t/cache_dir/*/*/*/*/*.bin
 // RUN: %{cache_vars} %{run-unfiltered-devices} %t.out 2>&1 | FileCheck %s --check-prefixes RESULT1
 // ******************************
 //
 // REDEFINE: %{build_cmd} = %clangxx -fsycl -fsycl-targets=spir64_gen -Xsycl-target-backend=spir64_gen "-device acm-g10" %s
 // ******************************
 // Check the logs first.
-// RUN: %{build_cmd} -DVALUE=1 -o %t.out
-// RUN: rm -rf %t/cache_dir/*
+// RUN: %{run-aux} %{build_cmd} -DVALUE=1 -o %t.out
+// RUN: %{run-aux} rm -rf %t/cache_dir/*
 // RUN: %{cache_vars} %{run-unfiltered-devices} %t.out 2>&1 | FileCheck %s  --check-prefixes=CHECK-CACHE-WRITE
 // RUN: %{cache_vars} %{run-unfiltered-devices} %t.out 2>&1 | FileCheck %s  --check-prefixes=CHECK-CACHE-READ
 // RUN: %{cache_vars} %{run-unfiltered-devices} %t.out 2>&1 | FileCheck %s  --check-prefixes=CHECK-CACHE-READ
@@ -40,30 +39,30 @@
 //
 // Now try to substitute the cached image and verify it is actually taken and
 // the code/binary there is executed.
-// RUN: mv %t/cache_dir/*/*/*/*/*.bin %t.value1.bin
-// RUN: rm -rf %t/cache_dir/*
-// RUN: %{build_cmd} -DVALUE=2 -o %t.out
+// RUN: %{run-aux} mv %t/cache_dir/*/*/*/*/*.bin %t.value1.bin
+// RUN: %{run-aux} rm -rf %t/cache_dir/*
+// RUN: %{run-aux} %{build_cmd} -DVALUE=2 -o %t.out
 // RUN: %{cache_vars} %{run-unfiltered-devices} %t.out 2>&1 | FileCheck %s --check-prefixes RESULT2
-// RUN: mv %t.value1.bin %t/cache_dir/*/*/*/*/*.bin
+// RUN: %{run-aux} mv %t.value1.bin %t/cache_dir/*/*/*/*/*.bin
 // RUN: %{cache_vars} %{run-unfiltered-devices} %t.out 2>&1 | FileCheck %s --check-prefixes RESULT1
 // ******************************
 //
 // REDEFINE: %{build_cmd} = %clangxx -fsycl -fsycl-targets=spir64_gen -Xsycl-target-backend=spir64_gen "-device acm-g10,acm-g11" %s
 // ******************************
 // Check the logs first.
-// RUN: %{build_cmd} -DVALUE=1 -o %t.out
-// RUN: rm -rf %t/cache_dir/*
+// RUN: %{run-aux} %{build_cmd} -DVALUE=1 -o %t.out
+// RUN: %{run-aux} rm -rf %t/cache_dir/*
 // RUN: %{cache_vars} %{run-unfiltered-devices} %t.out 2>&1 | FileCheck %s  --check-prefixes=CHECK-CACHE-WRITE
 // RUN: %{cache_vars} %{run-unfiltered-devices} %t.out 2>&1 | FileCheck %s  --check-prefixes=CHECK-CACHE-READ
 // RUN: %{cache_vars} %{run-unfiltered-devices} %t.out 2>&1 | FileCheck %s  --check-prefixes=CHECK-CACHE-READ
 //
 // Now try to substitute the cached image and verify it is actually taken and
 // the code/binary there is executed.
-// RUN: mv %t/cache_dir/*/*/*/*/*.bin %t.value1.bin
-// RUN: rm -rf %t/cache_dir/*
-// RUN: %{build_cmd} -DVALUE=2 -o %t.out
+// RUN: %{run-aux} mv %t/cache_dir/*/*/*/*/*.bin %t.value1.bin
+// RUN: %{run-aux} rm -rf %t/cache_dir/*
+// RUN: %{run-aux} %{build_cmd} -DVALUE=2 -o %t.out
 // RUN: %{cache_vars} %{run-unfiltered-devices} %t.out 2>&1 | FileCheck %s --check-prefixes RESULT2
-// RUN: mv %t.value1.bin %t/cache_dir/*/*/*/*/*.bin
+// RUN: %{run-aux} mv %t.value1.bin %t/cache_dir/*/*/*/*/*.bin
 // RUN: %{cache_vars} %{run-unfiltered-devices} %t.out 2>&1 | FileCheck %s --check-prefixes RESULT1
 // ******************************
 
diff --git a/sycl/test-e2e/NewOffloadDriver/aot-cpu.cpp b/sycl/test-e2e/NewOffloadDriver/aot-cpu.cpp
index 0016e8498bd9d..8926780d1912a 100644
--- a/sycl/test-e2e/NewOffloadDriver/aot-cpu.cpp
+++ b/sycl/test-e2e/NewOffloadDriver/aot-cpu.cpp
@@ -1,11 +1,11 @@
 // REQUIRES: opencl-aot, cpu
-// REQUIRES: build-and-run-mode
 
+// CPU AOT targets host isa, so we compile on the run system instead.
 // Test with `--offload-new-driver`
-// RUN: %clangxx -fsycl -fsycl-device-code-split=per_source -fsycl-targets=spir64_x86_64 -I %S/Inputs -o %t.out %S/split-per-source-main.cpp %S/Inputs/split-per-source-second-file.cpp \
+// RUN: %{run-aux} %clangxx -fsycl -fsycl-device-code-split=per_source -fsycl-targets=spir64_x86_64 -I %S/Inputs -o %t.out %S/split-per-source-main.cpp %S/Inputs/split-per-source-second-file.cpp \
 // RUN: -fsycl-dead-args-optimization --offload-new-driver
 // RUN: %{run} %t.out
 
 // Test -O0 with `--offload-new-driver`
-// RUN: %clangxx %O0 -fsycl -fsycl-targets=spir64-x86_64 %S/Inputs/aot.cpp
+// RUN: %{run-aux} %clangxx %O0 -fsycl -fsycl-targets=spir64-x86_64 %S/Inputs/aot.cpp
 // RUN: %{run} %t.out
diff --git a/sycl/test-e2e/NewOffloadDriver/cpu.cpp b/sycl/test-e2e/NewOffloadDriver/cpu.cpp
index 943ea71c3386e..4e61ff88ad87d 100644
--- a/sycl/test-e2e/NewOffloadDriver/cpu.cpp
+++ b/sycl/test-e2e/NewOffloadDriver/cpu.cpp
@@ -7,10 +7,10 @@
 //===---------------------------------------------------------------------===//
 
 // REQUIRES: opencl-aot, cpu
-// REQUIRES: build-and-run-mode
 
+// CPU AOT targets host isa, so we compile on the run system instead.
 // Test with  `--offload-new-driver`
-// RUN: %clangxx -fsycl -fsycl-targets=spir64_x86_64 --offload-new-driver %S/Inputs/aot.cpp -o %t.out
+// RUN: %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 --offload-new-driver %S/Inputs/aot.cpp -o %t.out
 // RUN: %{run} %t.out
 
 // Test that opencl-aot can handle multiple build options.
diff --git a/sycl/test-e2e/NonUniformGroups/ballot_group.cpp b/sycl/test-e2e/NonUniformGroups/ballot_group.cpp
index 1532c54879ecc..3824edaf4af89 100644
--- a/sycl/test-e2e/NonUniformGroups/ballot_group.cpp
+++ b/sycl/test-e2e/NonUniformGroups/ballot_group.cpp
@@ -1,10 +1,10 @@
 // RUN: %{build} -o %t.out
 // RUN: %{run} %t.out
 //
-// RUN: %if any-device-is-cpu && opencl-aot %{ %clangxx -fsycl -fsycl-targets=spir64_x86_64 -o %t.x86.out %s %}
+// CPU AOT targets host isa, so we compile on the run system instead.
+// RUN: %if any-device-is-cpu && opencl-aot %{ %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 -o %t.x86.out %s %}
 // RUN: %if cpu %{ %{run} %t.x86.out %}
 //
-// REQUIRES: build-and-run-mode
 // REQUIRES: cpu || gpu
 // UNSUPPORTED: hip
 
diff --git a/sycl/test-e2e/NonUniformGroups/ballot_group_algorithms.cpp b/sycl/test-e2e/NonUniformGroups/ballot_group_algorithms.cpp
index da9154ae602d0..6c6cda9708530 100644
--- a/sycl/test-e2e/NonUniformGroups/ballot_group_algorithms.cpp
+++ b/sycl/test-e2e/NonUniformGroups/ballot_group_algorithms.cpp
@@ -1,10 +1,10 @@
 // RUN: %{build} -o %t.out
 // RUN: %{run} %t.out
 //
-// RUN: %if any-device-is-cpu && opencl-aot %{ %clangxx -fsycl -fsycl-targets=spir64_x86_64 -o %t.x86.out %s %}
+// CPU AOT targets host isa, so we compile on the run system instead.
+// RUN: %if any-device-is-cpu && opencl-aot %{ %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 -o %t.x86.out %s %}
 // RUN: %if cpu %{ %{run} %t.x86.out %}
 //
-// REQUIRES: build-and-run-mode
 // REQUIRES: cpu || gpu
 // REQUIRES: sg-32
 // REQUIRES: aspect-ext_oneapi_ballot_group
diff --git a/sycl/test-e2e/NonUniformGroups/fixed_size_group.cpp b/sycl/test-e2e/NonUniformGroups/fixed_size_group.cpp
index 5d2b98b58ddc2..95533b3651bb1 100644
--- a/sycl/test-e2e/NonUniformGroups/fixed_size_group.cpp
+++ b/sycl/test-e2e/NonUniformGroups/fixed_size_group.cpp
@@ -1,10 +1,10 @@
 // RUN: %{build} -o %t.out
 // RUN: %{run} %t.out
 //
-// RUN: %if any-device-is-cpu && opencl-aot %{ %clangxx -fsycl -fsycl-targets=spir64_x86_64 -o %t.x86.out %s %}
+// CPU AOT targets host isa, so we compile on the run system instead.
+// RUN: %if any-device-is-cpu && opencl-aot %{ %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 -o %t.x86.out %s %}
 // RUN: %if cpu %{ %{run} %t.x86.out %}
 //
-// REQUIRES: build-and-run-mode
 // REQUIRES: cpu || gpu
 // UNSUPPORTED: hip
 // REQUIRES: sg-32
diff --git a/sycl/test-e2e/NonUniformGroups/fixed_size_group_algorithms.cpp b/sycl/test-e2e/NonUniformGroups/fixed_size_group_algorithms.cpp
index 07898d0146fb9..afd72070cd7c9 100644
--- a/sycl/test-e2e/NonUniformGroups/fixed_size_group_algorithms.cpp
+++ b/sycl/test-e2e/NonUniformGroups/fixed_size_group_algorithms.cpp
@@ -1,13 +1,15 @@
 // RUN: %{build} -fsycl-device-code-split=per_kernel -o %t.out
 // RUN: %{run} %t.out
 //
-// RUN: %if any-device-is-cpu && opencl-aot %{ %clangxx -fsycl -fsycl-targets=spir64_x86_64 -fsycl-device-code-split=per_kernel -o %t.x86.out %s %}
+// CPU AOT targets host isa, so we compile on the run system instead.
+// RUN: %if any-device-is-cpu && opencl-aot %{ %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 -fsycl-device-code-split=per_kernel -o %t.x86.out %s %}
 // RUN: %if cpu %{ %{run} %t.x86.out %}
 //
-// REQUIRES: build-and-run-mode
 // REQUIRES: cpu || gpu
 // REQUIRES: sg-32
 // REQUIRES: aspect-ext_oneapi_fixed_size_group
+// UNSUPPORTED: target-amd
+// UNSUPPORTED-INTENDED: fixed_size_group aspect not available on amd
 
 // UNSUPPORTED: spirv-backend
 // UNSUPPORTED-TRACKER: CMPLRLLVM-64702
diff --git a/sycl/test-e2e/NonUniformGroups/opportunistic_group.cpp b/sycl/test-e2e/NonUniformGroups/opportunistic_group.cpp
index 2069113c3f939..cca46c7dd785d 100644
--- a/sycl/test-e2e/NonUniformGroups/opportunistic_group.cpp
+++ b/sycl/test-e2e/NonUniformGroups/opportunistic_group.cpp
@@ -1,10 +1,10 @@
 // RUN: %{build} -o %t.out
 // RUN: %{run} %t.out
 //
-// RUN: %if any-device-is-cpu && opencl-aot %{ %clangxx -fsycl -fsycl-targets=spir64_x86_64 -o %t.x86.out %s %}
+// CPU AOT targets host isa, so we compile on the run system instead.
+// RUN: %if any-device-is-cpu && opencl-aot %{ %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 -o %t.x86.out %s %}
 // RUN: %if cpu %{ %{run} %t.x86.out %}
 //
-// REQUIRES: build-and-run-mode
 // REQUIRES: cpu || gpu
 // UNSUPPORTED: hip
 
diff --git a/sycl/test-e2e/NonUniformGroups/opportunistic_group_algorithms.cpp b/sycl/test-e2e/NonUniformGroups/opportunistic_group_algorithms.cpp
index 8e7626650e01c..7e20c1bfeccfc 100644
--- a/sycl/test-e2e/NonUniformGroups/opportunistic_group_algorithms.cpp
+++ b/sycl/test-e2e/NonUniformGroups/opportunistic_group_algorithms.cpp
@@ -1,10 +1,10 @@
 // RUN: %{build} -o %t.out
 // RUN: %{run} %t.out
 //
-// RUN: %if any-device-is-cpu && opencl-aot %{ %clangxx -fsycl -fsycl-targets=spir64_x86_64 -o %t.x86.out %s %}
+// CPU AOT targets host isa, so we compile on the run system instead.
+// RUN: %if any-device-is-cpu && opencl-aot %{ %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 -o %t.x86.out %s %}
 // RUN: %if cpu %{ %{run} %t.x86.out %}
 //
-// REQUIRES: build-and-run-mode
 // REQUIRES: cpu || gpu
 // REQUIRES: sg-32
 // REQUIRES: aspect-ext_oneapi_opportunistic_group
diff --git a/sycl/test-e2e/NonUniformGroups/tangle_group.cpp b/sycl/test-e2e/NonUniformGroups/tangle_group.cpp
index 686d1b34c6185..bc4a3e4a35a22 100644
--- a/sycl/test-e2e/NonUniformGroups/tangle_group.cpp
+++ b/sycl/test-e2e/NonUniformGroups/tangle_group.cpp
@@ -1,12 +1,12 @@
 // RUN: %{build} -fno-sycl-early-optimizations -o %t.out
 // RUN: %{run} %t.out
 //
-// RUN: %if any-device-is-cpu && opencl-aot %{ %clangxx -fsycl -fsycl-targets=spir64_x86_64 -fno-sycl-early-optimizations -o %t.x86.out %s %}
+// CPU AOT targets host isa, so we compile on the run system instead.
+// RUN: %if any-device-is-cpu && opencl-aot %{ %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 -fno-sycl-early-optimizations -o %t.x86.out %s %}
 // RUN: %if cpu %{ %{run} %t.x86.out %}
 //
-// REQUIRES: build-and-run-mode
 // REQUIRES: cpu || gpu
-// UNSUPPORTED: cuda || hip
+// UNSUPPORTED: target-nvidia || target-amd
 
 #include <sycl/detail/core.hpp>
 #include <sycl/ext/oneapi/experimental/tangle_group.hpp>
diff --git a/sycl/test-e2e/NonUniformGroups/tangle_group_algorithms.cpp b/sycl/test-e2e/NonUniformGroups/tangle_group_algorithms.cpp
index 51070ed1731e4..f49ce691d2c75 100644
--- a/sycl/test-e2e/NonUniformGroups/tangle_group_algorithms.cpp
+++ b/sycl/test-e2e/NonUniformGroups/tangle_group_algorithms.cpp
@@ -1,14 +1,16 @@
 // RUN: %{build} -fno-sycl-early-optimizations -o %t.out
 // RUN: %{run} %t.out
 //
-// RUN: %if any-device-is-cpu && opencl-aot %{ %clangxx -fsycl -fsycl-targets=spir64_x86_64 -fno-sycl-early-optimizations -o %t.x86.out %s %}
+// CPU AOT targets host isa, so we compile on the run system instead.
+// RUN: %if any-device-is-cpu && opencl-aot %{ %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 -fno-sycl-early-optimizations -o %t.x86.out %s %}
 // RUN: %if cpu %{ %{run} %t.x86.out %}
 //
-// REQUIRES: build-and-run-mode
 // REQUIRES: cpu || gpu
 // REQUIRES: sg-32
 // REQUIRES: aspect-ext_oneapi_tangle_group
-// UNSUPPORTED: cuda || windows
+// UNSUPPORTED: target-amd
+// UNSUPPORTED-INTENDED: tangle groups not available on amd
+// UNSUPPORTED: target-nvidia || windows
 // Tangle groups exhibit unpredictable behavior on Windows.
 // The test is disabled while we investigate the root cause.
 
diff --git a/sycl/test-e2e/ProgramManager/multi_device_bundle/device_libs_and_caching.cpp b/sycl/test-e2e/ProgramManager/multi_device_bundle/device_libs_and_caching.cpp
index 877a78fae2e21..13e96de7e7244 100644
--- a/sycl/test-e2e/ProgramManager/multi_device_bundle/device_libs_and_caching.cpp
+++ b/sycl/test-e2e/ProgramManager/multi_device_bundle/device_libs_and_caching.cpp
@@ -1,5 +1,4 @@
-// REQUIRES: ocloc && gpu && linux && (opencl || level_zero)
-// REQUIRES: build-and-run-mode
+// REQUIRES: ocloc && gpu && linux && target-spir
 
 // Test to check several use cases for multi-device kernel bundles.
 // Test covers AOT and JIT cases. Kernel is using some math functions to enforce
@@ -12,22 +11,22 @@
 // manager can handle this as well. With this option program manager will
 // compile the main program, load and compile device libraries and then link
 // everything together.
-// RUN: %{build} -fsycl-device-lib-jit-link -o %t.out
+// RUN: %{build} -fsycl-device-lib-jit-link -o %t_jit.out
 
 // Check the default case when in-memory caching is enabled.
-// RUN: env NEOReadDebugKeys=1 CreateMultipleRootDevices=4 SYCL_UR_TRACE=2 %{run} %t.out | FileCheck %s --check-prefixes=CHECK-SPIRV-JIT-LINK-TRACE
+// RUN: env NEOReadDebugKeys=1 CreateMultipleRootDevices=4 SYCL_UR_TRACE=2 %{run} %t_jit.out | FileCheck %s --check-prefixes=CHECK-SPIRV-JIT-LINK-TRACE
 
 // Check the case when in-memory caching of the programs is disabled.
-// RUN: env SYCL_CACHE_IN_MEM=0 NEOReadDebugKeys=1 CreateMultipleRootDevices=4 %{run} %t.out
+// RUN: env SYCL_CACHE_IN_MEM=0 NEOReadDebugKeys=1 CreateMultipleRootDevices=4 %{run} %t_jit.out
 
 // Test AOT next.
-// RUN: %{build} -fsycl-targets=spir64_gen -Xsycl-target-backend=spir64_gen "-device *" -o %t.out
+// RUN: %{build} -fsycl-targets=spir64_gen -Xsycl-target-backend=spir64_gen "-device *" -o %t_aot.out
 
 // Check the default case when in-memory caching is enabled.
-// RUN: env NEOReadDebugKeys=1 CreateMultipleRootDevices=4 SYCL_UR_TRACE=2 %{run} %t.out | FileCheck %s --check-prefixes=CHECK-AOT-TRACE
+// RUN: env NEOReadDebugKeys=1 CreateMultipleRootDevices=4 SYCL_UR_TRACE=2 %{run} %t_aot.out | FileCheck %s --check-prefixes=CHECK-AOT-TRACE
 
 // Check the case when in-memory caching of the programs is disabled.
-// RUN: env SYCL_CACHE_IN_MEM=0 NEOReadDebugKeys=1 CreateMultipleRootDevices=4 %{run} %t.out
+// RUN: env SYCL_CACHE_IN_MEM=0 NEOReadDebugKeys=1 CreateMultipleRootDevices=4 %{run} %t_aot.out
 
 // Depends on SPIR-V Backend & run-time drivers version.
 // XFAIL: spirv-backend && run-mode
diff --git a/sycl/test-e2e/SpecConstants/2020/native_specialization_constant.cpp b/sycl/test-e2e/SpecConstants/2020/native_specialization_constant.cpp
index c15c4ba892ed7..da5101b277e31 100644
--- a/sycl/test-e2e/SpecConstants/2020/native_specialization_constant.cpp
+++ b/sycl/test-e2e/SpecConstants/2020/native_specialization_constant.cpp
@@ -4,7 +4,8 @@
 
 // FIXME: This set is never satisfied all at once in our infrastructure.
 // REQUIRES: opencl, level-zero, cpu, gpu, opencl-aot, ocloc
-// REQUIRES: build-and-run-mode
+// XFAIL: *
+// XFAIL-TRACKER: https://github.com/intel/llvm/issues/17107
 
 // RUN: %clangxx -fsycl -DJIT %s -o %t.out
 // RUN: %{run} %t.out
diff --git a/sycl/test-e2e/SpecConstants/2020/non_native/SpecConstBuffer.cpp b/sycl/test-e2e/SpecConstants/2020/non_native/SpecConstBuffer.cpp
index bc7145049ddcb..7c2076e3ee77e 100644
--- a/sycl/test-e2e/SpecConstants/2020/non_native/SpecConstBuffer.cpp
+++ b/sycl/test-e2e/SpecConstants/2020/non_native/SpecConstBuffer.cpp
@@ -1,7 +1,7 @@
 // REQUIRES: opencl-aot, cpu
-// REQUIRES: build-and-run-mode
 
-// RUN: %clangxx -fsycl -fsycl-targets=spir64_x86_64 %S/Inputs/common.cpp -o %t.out \
+// CPU AOT targets host isa, so we compile on the run system instead.
+// RUN: %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 %S/Inputs/common.cpp -o %t.out \
 // RUN:          -fsycl-dead-args-optimization
 // RUN: env SYCL_UR_TRACE=2 %{run} %t.out | FileCheck %s
 
diff --git a/sycl/test-e2e/SpecConstants/2020/non_native/cpu.cpp b/sycl/test-e2e/SpecConstants/2020/non_native/cpu.cpp
index f6b7b836e873c..6fed8faa7fbd8 100644
--- a/sycl/test-e2e/SpecConstants/2020/non_native/cpu.cpp
+++ b/sycl/test-e2e/SpecConstants/2020/non_native/cpu.cpp
@@ -1,7 +1,7 @@
 // REQUIRES: opencl-aot, cpu
-// REQUIRES: build-and-run-mode
 
-// RUN: %clangxx -fsycl -fsycl-targets=spir64_x86_64 %S/Inputs/common.cpp -o %t.out \
+// CPU AOT targets host isa, so we compile on the run system instead.
+// RUN: %{run-aux} %clangxx -fsycl -fsycl-targets=spir64_x86_64 %S/Inputs/common.cpp -o %t.out \
 // RUN:          -fsycl-dead-args-optimization
 // RUN: %{run} %t.out
 
diff --git a/sycl/test-e2e/bindless_images/cubemap/cubemap_sampled.cpp b/sycl/test-e2e/bindless_images/cubemap/cubemap_sampled.cpp
index 23051ec0850ef..e4a5c660417ba 100644
--- a/sycl/test-e2e/bindless_images/cubemap/cubemap_sampled.cpp
+++ b/sycl/test-e2e/bindless_images/cubemap/cubemap_sampled.cpp
@@ -1,6 +1,6 @@
+// REQUIRES: target-nvidia
 // REQUIRES: aspect-ext_oneapi_cubemap
 // REQUIRES: aspect-ext_oneapi_cubemap_seamless_filtering
-// REQUIRES: build-and-run-mode
 
 // RUN: %{build} -o %t.out
 // RUN: %{run-unfiltered-devices} %t.out
diff --git a/sycl/test-e2e/bindless_images/cubemap/cubemap_unsampled.cpp b/sycl/test-e2e/bindless_images/cubemap/cubemap_unsampled.cpp
index 80ad06ab738d2..5f2ef75faafd1 100644
--- a/sycl/test-e2e/bindless_images/cubemap/cubemap_unsampled.cpp
+++ b/sycl/test-e2e/bindless_images/cubemap/cubemap_unsampled.cpp
@@ -1,5 +1,4 @@
 // REQUIRES: aspect-ext_oneapi_cubemap
-// REQUIRES: build-and-run-mode
 
 // RUN: %{build} -o %t.out
 // RUN: %{run-unfiltered-devices} %t.out
diff --git a/sycl/test-e2e/bindless_images/examples/example_5_sample_cubemap.cpp b/sycl/test-e2e/bindless_images/examples/example_5_sample_cubemap.cpp
index 3e585eba2be9f..0f93a5b9d2231 100644
--- a/sycl/test-e2e/bindless_images/examples/example_5_sample_cubemap.cpp
+++ b/sycl/test-e2e/bindless_images/examples/example_5_sample_cubemap.cpp
@@ -1,5 +1,4 @@
-// REQUIRES: aspect-ext_oneapi_cubemap
-// REQUIRES: build-and-run-mode
+// REQUIRES: target-nvidia, aspect-ext_oneapi_cubemap
 
 // RUN: %{build} -o %t.out
 // RUN: %{run-unfiltered-devices} %t.out
diff --git a/sycl/test-e2e/format.py b/sycl/test-e2e/format.py
index c8f5dbd4e9815..3b9ba2b174ac2 100644
--- a/sycl/test-e2e/format.py
+++ b/sycl/test-e2e/format.py
@@ -309,13 +309,8 @@ def get_extra_env(sycl_devices):
                 for i in ["%{run}", "%{run-unfiltered-devices}", "%{run-aux}"]
             )
 
-            ignore_line_filtering = (
-                "build-and-run-mode" in test.requires
-                and test.config.fallback_build_run_only
-            )
-            if not ignore_line_filtering and (
-                (is_run_line and test.config.test_mode == "build-only")
-                or (not is_run_line and test.config.test_mode == "run-only")
+            if (is_run_line and test.config.test_mode == "build-only") or (
+                not is_run_line and test.config.test_mode == "run-only"
             ):
                 continue
 
diff --git a/sycl/test-e2e/lit.cfg.py b/sycl/test-e2e/lit.cfg.py
index 40851306068a7..43d2bc012d4ae 100644
--- a/sycl/test-e2e/lit.cfg.py
+++ b/sycl/test-e2e/lit.cfg.py
@@ -58,20 +58,14 @@
 config.test_mode = lit_config.params.get("test-mode", "full")
 config.fallback_build_run_only = False
 if config.test_mode == "full":
+    config.available_features.add("build-mode")
     config.available_features.add("run-mode")
-    config.available_features.add("build-and-run-mode")
 elif config.test_mode == "run-only":
     lit_config.note("run-only test mode enabled, only executing tests")
-    # run-only uses external shell, some tests might have hacks to workaround
-    # failures caused by that.
-    config.available_features.add("test-mode-run-only")
-
     config.available_features.add("run-mode")
-    if lit_config.params.get("fallback-to-build-if-requires-build-and-run", False):
-        config.available_features.add("build-and-run-mode")
-        config.fallback_build_run_only = True
 elif config.test_mode == "build-only":
     lit_config.note("build-only test mode enabled, only compiling tests")
+    config.available_features.add("build-mode")
     config.sycl_devices = []
     if not config.amd_arch:
         config.amd_arch = "gfx1030"