Use a single workspace for flashinfer (#1077)

benchmark/gsm8k/bench_sglang.py

@@ -64,7 +64,7 @@ def main(args):
     @sgl.function
     def few_shot_gsm8k(s, question):
         s += few_shot_examples + question
-        s += sgl.gen("answer", max_tokens=512, stop="Question")
+        s += sgl.gen("answer", max_tokens=512, stop=["Question", "Assistant:"])
 
     #####################################
     ########## SGL Program End ##########

python/sglang/global_config.py

@@ -27,7 +27,7 @@ class GlobalConfig:
         # Runtime constants: others
         self.num_continue_decode_steps = 10
         self.retract_decode_steps = 20
-        self.flashinfer_workspace_size = 192 * 1024 * 1024
+        self.flashinfer_workspace_size = 384 * 1024 * 1024
 
         # Output tokenization configs
         self.skip_special_tokens_in_output = True

python/sglang/srt/model_executor/cuda_graph_runner.py

@@ -120,13 +120,13 @@ class CudaGraphRunner:
         )
         if model_runner.sliding_window_size is None:
             self.flashinfer_workspace_buffer = (
-                self.model_runner.flashinfer_workspace_buffers[0]
+                self.model_runner.flashinfer_workspace_buffer
             )
         else:
-            self.flashinfer_workspace_buffers = [
+            self.flashinfer_workspace_buffer = (
-                self.model_runner.flashinfer_workspace_buffers[0],
+                self.model_runner.flashinfer_workspace_buffer
-                self.model_runner.flashinfer_workspace_buffers[2],
+            )
-            ]
+
             self.flashinfer_kv_indptr = [
                 self.flashinfer_kv_indptr,
                 self.flashinfer_kv_indptr.clone(),
@@ -200,7 +200,7 @@ class CudaGraphRunner:
             for i in range(2):
                 flashinfer_decode_wrapper.append(
                     BatchDecodeWithPagedKVCacheWrapper(
-                        self.flashinfer_workspace_buffers[i],
+                        self.flashinfer_workspace_buffer,
                         "NHD",
                         use_cuda_graph=True,
                         use_tensor_cores=use_tensor_cores,

python/sglang/srt/model_executor/model_runner.py

@@ -318,28 +318,26 @@ class ModelRunner:
             use_tensor_cores = False
 
         if self.sliding_window_size is None:
-            self.flashinfer_workspace_buffers = torch.empty(
+            self.flashinfer_workspace_buffer = torch.empty(
-                2,
                 global_config.flashinfer_workspace_size,
                 dtype=torch.uint8,
                 device="cuda",
             )
             self.flashinfer_prefill_wrapper_ragged = (
                 BatchPrefillWithRaggedKVCacheWrapper(
-                    self.flashinfer_workspace_buffers[0], "NHD"
+                    self.flashinfer_workspace_buffer, "NHD"
                 )
             )
             self.flashinfer_prefill_wrapper_paged = BatchPrefillWithPagedKVCacheWrapper(
-                self.flashinfer_workspace_buffers[1], "NHD"
+                self.flashinfer_workspace_buffer, "NHD"
             )
             self.flashinfer_decode_wrapper = BatchDecodeWithPagedKVCacheWrapper(
-                self.flashinfer_workspace_buffers[0],
+                self.flashinfer_workspace_buffer,
                 "NHD",
                 use_tensor_cores=use_tensor_cores,
             )
         else:
             self.flashinfer_workspace_buffers = torch.empty(
-                4,
                 global_config.flashinfer_workspace_size,
                 dtype=torch.uint8,
                 device="cuda",
@@ -350,17 +348,17 @@ class ModelRunner:
             for i in range(2):
                 self.flashinfer_prefill_wrapper_ragged.append(
                     BatchPrefillWithRaggedKVCacheWrapper(
-                        self.flashinfer_workspace_buffers[2 * i + 0], "NHD"
+                        self.flashinfer_workspace_buffer, "NHD"
                     )
                 )
                 self.flashinfer_prefill_wrapper_paged.append(
                     BatchPrefillWithPagedKVCacheWrapper(
-                        self.flashinfer_workspace_buffers[2 * i + 1], "NHD"
+                        self.flashinfer_workspace_buffer, "NHD"
                     )
                 )
                 self.flashinfer_decode_wrapper.append(
                     BatchDecodeWithPagedKVCacheWrapper(
-                        self.flashinfer_workspace_buffers[2 * i + 0],
+                        self.flashinfer_workspace_buffer,
                         "NHD",
                         use_tensor_cores=use_tensor_cores,
                     )

python/sglang/srt/server.py

@@ -381,7 +381,7 @@ def _set_envs_and_config(server_args: ServerArgs):
     if not server_args.disable_flashinfer:
         assert_pkg_version(
             "flashinfer",
-            "0.1.4",
+            "0.1.5",
             "Please uninstall the old version and "
             "reinstall the latest version by following the instructions "
             "at https://docs.flashinfer.ai/installation.html.",