diff --git a/Cargo.lock b/Cargo.lock
index 5ca8fea3..4092195f 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -540,6 +540,15 @@ version = "1.0.97"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "dcfed56ad506cb2c684a14971b8861fdc3baaaae314b9e5f9bb532cbe3ba7a4f"
 
+[[package]]
+name = "approx"
+version = "0.5.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "cab112f0a86d568ea0e627cc1d6be74a1e9cd55214684db5561995f6dad897c6"
+dependencies = [
+ "num-traits",
+]
+
 [[package]]
 name = "ark-ff"
 version = "0.3.0"
@@ -798,6 +807,12 @@ dependencies = [
  "syn 2.0.99",
 ]
 
+[[package]]
+name = "bitflags"
+version = "1.3.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "bef38d45163c2f1dde094a7dfd33ccf595c92905c8f8f4fdc18d06fb1037718a"
+
 [[package]]
 name = "bitflags"
 version = "2.9.0"
@@ -971,6 +986,20 @@ name = "bytemuck"
 version = "1.22.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "b6b1fc10dbac614ebc03540c9dbd60e83887fda27794998c6528f1782047d540"
+dependencies = [
+ "bytemuck_derive",
+]
+
+[[package]]
+name = "bytemuck_derive"
+version = "1.8.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3fa76293b4f7bb636ab88fd78228235b5248b4d05cc589aed610f954af5d7c7a"
+dependencies = [
+ "proc-macro2",
+ "quote",
+ "syn 2.0.99",
+]
 
 [[package]]
 name = "byteorder"
@@ -1316,6 +1345,52 @@ dependencies = [
  "typenum",
 ]
 
+[[package]]
+name = "cust"
+version = "0.3.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0d6cc71911e179f12483b9734120b45bd00bf64fab085cc4818428523eedd469"
+dependencies = [
+ "bitflags 1.3.2",
+ "bytemuck",
+ "cust_core",
+ "cust_derive",
+ "cust_raw",
+ "find_cuda_helper",
+]
+
+[[package]]
+name = "cust_core"
+version = "0.1.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "039f79662cb8f890cbf335e818cd522d6e3a53fe63f61d1aaaf859cd3d975f06"
+dependencies = [
+ "cust_derive",
+ "glam",
+ "mint",
+ "vek",
+]
+
+[[package]]
+name = "cust_derive"
+version = "0.2.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e8a3bc95fe629aed92b2423de6ccff9e40174b21d19cb6ee6281a4d04ac72f66"
+dependencies = [
+ "proc-macro2",
+ "quote",
+ "syn 1.0.109",
+]
+
+[[package]]
+name = "cust_raw"
+version = "0.11.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "fbf40d6ade12cb9828bbc844b9875c7b93d25e67a3c9bf61c7aa3ae09e402bf8"
+dependencies = [
+ "find_cuda_helper",
+]
+
 [[package]]
 name = "darling"
 version = "0.20.10"
@@ -1460,6 +1535,15 @@ dependencies = [
  "subtle",
 ]
 
+[[package]]
+name = "directories"
+version = "5.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "9a49173b84e034382284f27f1af4dcbbd231ffa358c0fe316541a7337f376a35"
+dependencies = [
+ "dirs-sys",
+]
+
 [[package]]
 name = "dirs"
 version = "5.0.1"
@@ -1700,6 +1784,15 @@ dependencies = [
  "syn 1.0.109",
 ]
 
+[[package]]
+name = "find_cuda_helper"
+version = "0.2.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f9f9e65c593dd01ac77daad909ea4ad17f0d6d1776193fc8ea766356177abdad"
+dependencies = [
+ "glob",
+]
+
 [[package]]
 name = "fixed-hash"
 version = "0.8.0"
@@ -1860,6 +1953,15 @@ dependencies = [
  "syn 2.0.99",
 ]
 
+[[package]]
+name = "glam"
+version = "0.20.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f43e957e744be03f5801a55472f593d43fabdebf25a4585db250f04d86b1675f"
+dependencies = [
+ "num-traits",
+]
+
 [[package]]
 name = "glob"
 version = "0.3.2"
@@ -2139,6 +2241,15 @@ dependencies = [
  "digest 0.10.7",
 ]
 
+[[package]]
+name = "home"
+version = "0.5.11"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "589533453244b0995c858700322199b2becb13b627df2851f64a2775d024abcf"
+dependencies = [
+ "windows-sys 0.59.0",
+]
+
 [[package]]
 name = "iana-time-zone"
 version = "0.1.61"
@@ -2496,7 +2607,7 @@ version = "0.1.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "c0ff37bd590ca25063e35af745c343cb7a0271906fb7b37e4813e8f79f00268d"
 dependencies = [
- "bitflags",
+ "bitflags 2.9.0",
  "libc",
 ]
 
@@ -2512,6 +2623,16 @@ version = "0.7.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "23fb14cb19457329c82206317a5663005a4d404783dc74f4252769b0d5f42856"
 
+[[package]]
+name = "lock_api"
+version = "0.4.12"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "07af8b9cdd281b7915f413fa73f29ebd5d55d0d3f0155584dade1ff18cea1b17"
+dependencies = [
+ "autocfg",
+ "scopeguard",
+]
+
 [[package]]
 name = "lockfree-object-pool"
 version = "0.1.6"
@@ -2539,6 +2660,16 @@ dependencies = [
  "regex-automata 0.1.10",
 ]
 
+[[package]]
+name = "matrixmultiply"
+version = "0.3.9"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "9380b911e3e96d10c1f415da0876389aaf1b56759054eeb0de7df940c456ba1a"
+dependencies = [
+ "autocfg",
+ "rawpointer",
+]
+
 [[package]]
 name = "maybe-rayon"
 version = "0.1.1"
@@ -2647,6 +2778,12 @@ dependencies = [
  "adler2",
 ]
 
+[[package]]
+name = "mint"
+version = "0.5.9"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e53debba6bda7a793e5f99b8dacf19e626084f525f7829104ba9898f367d85ff"
+
 [[package]]
 name = "modular-bitfield"
 version = "0.11.2"
@@ -2688,6 +2825,20 @@ dependencies = [
  "syn 2.0.99",
 ]
 
+[[package]]
+name = "ndarray"
+version = "0.15.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "adb12d4e967ec485a5f71c6311fe28158e9d6f4bc4a447b474184d0f91a8fa32"
+dependencies = [
+ "matrixmultiply",
+ "num-complex",
+ "num-integer",
+ "num-traits",
+ "rawpointer",
+ "rayon",
+]
+
 [[package]]
 name = "nibble_vec"
 version = "0.1.0"
@@ -2806,6 +2957,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "071dfc062690e90b734c0b2273ce72ad0ffa95f0c74596bc250dcfd960262841"
 dependencies = [
  "autocfg",
+ "libm",
 ]
 
 [[package]]
@@ -3714,7 +3866,7 @@ dependencies = [
 [[package]]
 name = "openvm-stark-backend"
 version = "1.0.0-rc.0"
-source = "git+https://github.com/openvm-org/stark-backend.git?rev=bc364134b8315c27bfd29c6e77ac79fe77090137#bc364134b8315c27bfd29c6e77ac79fe77090137"
+source = "git+ssh://git@github.com/scroll-tech/openvm-stark-gpu.git?branch=main#d587d1ce572fe291279c356dc836964ae8fc3843"
 dependencies = [
  "async-trait",
  "cfg-if",
@@ -3726,6 +3878,9 @@ dependencies = [
  "p3-challenger",
  "p3-commit",
  "p3-field",
+ "p3-gpu-backend",
+ "p3-gpu-base",
+ "p3-gpu-module",
  "p3-matrix",
  "p3-maybe-rayon",
  "p3-uni-stark",
@@ -3740,7 +3895,7 @@ dependencies = [
 [[package]]
 name = "openvm-stark-sdk"
 version = "1.0.0-rc.0"
-source = "git+https://github.com/openvm-org/stark-backend.git?rev=bc364134b8315c27bfd29c6e77ac79fe77090137#bc364134b8315c27bfd29c6e77ac79fe77090137"
+source = "git+ssh://git@github.com/scroll-tech/openvm-stark-gpu.git?branch=main#d587d1ce572fe291279c356dc836964ae8fc3843"
 dependencies = [
  "derivative",
  "derive_more 0.99.19",
@@ -3756,6 +3911,7 @@ dependencies = [
  "p3-dft",
  "p3-fri",
  "p3-goldilocks",
+ "p3-gpu-backend",
  "p3-keccak",
  "p3-merkle-tree",
  "p3-poseidon",
@@ -3825,7 +3981,7 @@ dependencies = [
 [[package]]
 name = "p3-air"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
  "p3-field",
  "p3-matrix",
@@ -3834,7 +3990,7 @@ dependencies = [
 [[package]]
 name = "p3-baby-bear"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
  "p3-field",
  "p3-mds",
@@ -3848,7 +4004,7 @@ dependencies = [
 [[package]]
 name = "p3-blake3"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
  "blake3",
  "p3-symmetric",
@@ -3858,7 +4014,7 @@ dependencies = [
 [[package]]
 name = "p3-bn254-fr"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
  "ff 0.13.0",
  "halo2curves",
@@ -3873,7 +4029,7 @@ dependencies = [
 [[package]]
 name = "p3-challenger"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
  "p3-field",
  "p3-maybe-rayon",
@@ -3885,12 +4041,14 @@ dependencies = [
 [[package]]
 name = "p3-commit"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
+ "anyhow",
  "itertools 0.13.0",
  "p3-challenger",
  "p3-dft",
  "p3-field",
+ "p3-gpu-base",
  "p3-matrix",
  "p3-util",
  "serde",
@@ -3899,7 +4057,7 @@ dependencies = [
 [[package]]
 name = "p3-dft"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
  "itertools 0.13.0",
  "p3-field",
@@ -3912,7 +4070,7 @@ dependencies = [
 [[package]]
 name = "p3-field"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
  "itertools 0.13.0",
  "num-bigint 0.4.6",
@@ -3929,26 +4087,31 @@ dependencies = [
 [[package]]
 name = "p3-fri"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
+ "anyhow",
  "itertools 0.13.0",
  "p3-challenger",
  "p3-commit",
  "p3-dft",
  "p3-field",
+ "p3-gpu-backend",
+ "p3-gpu-base",
  "p3-interpolation",
  "p3-matrix",
  "p3-maybe-rayon",
+ "p3-merkle-tree",
  "p3-util",
  "rand",
  "serde",
  "tracing",
+ "zkhash",
 ]
 
 [[package]]
 name = "p3-goldilocks"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
  "num-bigint 0.4.6",
  "p3-dft",
@@ -3962,10 +4125,95 @@ dependencies = [
  "serde",
 ]
 
+[[package]]
+name = "p3-gpu-backend"
+version = "0.1.0"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
+dependencies = [
+ "anyhow",
+ "bytemuck",
+ "cc",
+ "cust",
+ "cust_raw",
+ "ff 0.13.0",
+ "itertools 0.13.0",
+ "lazy_static",
+ "ndarray",
+ "once_cell",
+ "p3-baby-bear",
+ "p3-commit",
+ "p3-dft",
+ "p3-field",
+ "p3-gpu-base",
+ "p3-gpu-build",
+ "p3-interpolation",
+ "p3-matrix",
+ "p3-maybe-rayon",
+ "p3-poseidon2",
+ "p3-util",
+ "parking_lot",
+ "paste",
+ "rand",
+ "rand_core",
+ "rayon",
+ "serde",
+ "sppark",
+ "tracing",
+ "transpose",
+]
+
+[[package]]
+name = "p3-gpu-base"
+version = "0.1.0"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
+dependencies = [
+ "anyhow",
+ "bytemuck",
+ "cust",
+ "cust_raw",
+ "hex",
+ "lazy_static",
+ "p3-field",
+ "p3-matrix",
+ "parking_lot",
+ "tracing",
+]
+
+[[package]]
+name = "p3-gpu-build"
+version = "0.1.0"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
+dependencies = [
+ "cc",
+ "directories",
+ "hex",
+ "p3-gpu-field",
+ "sha2",
+ "tempfile",
+]
+
+[[package]]
+name = "p3-gpu-field"
+version = "0.1.0"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
+
+[[package]]
+name = "p3-gpu-module"
+version = "0.1.0"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
+dependencies = [
+ "itertools 0.13.0",
+ "p3-field",
+ "p3-gpu-base",
+ "p3-matrix",
+ "p3-util",
+ "tracing",
+]
+
 [[package]]
 name = "p3-interpolation"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
  "p3-field",
  "p3-matrix",
@@ -3976,7 +4224,7 @@ dependencies = [
 [[package]]
 name = "p3-keccak"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
  "itertools 0.13.0",
  "p3-field",
@@ -3988,21 +4236,23 @@ dependencies = [
 [[package]]
 name = "p3-keccak-air"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
  "p3-air",
  "p3-field",
+ "p3-gpu-backend",
  "p3-matrix",
  "p3-maybe-rayon",
  "p3-util",
  "rand",
  "tracing",
+ "zkhash",
 ]
 
 [[package]]
 name = "p3-matrix"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
  "itertools 0.13.0",
  "p3-field",
@@ -4017,7 +4267,7 @@ dependencies = [
 [[package]]
 name = "p3-maybe-rayon"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
  "rayon",
 ]
@@ -4025,7 +4275,7 @@ dependencies = [
 [[package]]
 name = "p3-mds"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
  "itertools 0.13.0",
  "p3-dft",
@@ -4039,11 +4289,13 @@ dependencies = [
 [[package]]
 name = "p3-merkle-tree"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
+ "anyhow",
  "itertools 0.13.0",
  "p3-commit",
  "p3-field",
+ "p3-gpu-base",
  "p3-matrix",
  "p3-maybe-rayon",
  "p3-symmetric",
@@ -4056,7 +4308,7 @@ dependencies = [
 [[package]]
 name = "p3-monty-31"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
  "itertools 0.13.0",
  "num-bigint 0.4.6",
@@ -4077,7 +4329,7 @@ dependencies = [
 [[package]]
 name = "p3-poseidon"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
  "p3-field",
  "p3-mds",
@@ -4088,7 +4340,7 @@ dependencies = [
 [[package]]
 name = "p3-poseidon2"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
  "gcd",
  "p3-field",
@@ -4100,7 +4352,7 @@ dependencies = [
 [[package]]
 name = "p3-poseidon2-air"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
  "p3-air",
  "p3-field",
@@ -4116,7 +4368,7 @@ dependencies = [
 [[package]]
 name = "p3-symmetric"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
  "itertools 0.13.0",
  "p3-field",
@@ -4126,7 +4378,7 @@ dependencies = [
 [[package]]
 name = "p3-uni-stark"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
  "itertools 0.13.0",
  "p3-air",
@@ -4144,7 +4396,7 @@ dependencies = [
 [[package]]
 name = "p3-util"
 version = "0.1.0"
-source = "git+https://github.com/Plonky3/Plonky3.git?rev=b0591e9#b0591e9b82d58d10f86359875b5d5fa96433b4cf"
+source = "git+ssh://git@github.com/scroll-tech/plonky3-gpu.git?branch=openvm-v2#342685a5353519a56d26abea43ac40415f4189b3"
 dependencies = [
  "serde",
 ]
@@ -4196,6 +4448,29 @@ dependencies = [
  "syn 2.0.99",
 ]
 
+[[package]]
+name = "parking_lot"
+version = "0.12.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f1bf18183cf54e8d6059647fc3063646a1801cf30896933ec2311622cc4b9a27"
+dependencies = [
+ "lock_api",
+ "parking_lot_core",
+]
+
+[[package]]
+name = "parking_lot_core"
+version = "0.9.10"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "1e401f977ab385c9e4e3ab30627d6f26d00e2c73eef317493c4ec6d468726cf8"
+dependencies = [
+ "cfg-if",
+ "libc",
+ "redox_syscall",
+ "smallvec",
+ "windows-targets 0.52.6",
+]
+
 [[package]]
 name = "pasta_curves"
 version = "0.4.1"
@@ -4391,7 +4666,7 @@ checksum = "14cae93065090804185d3b75f0bf93b8eeda30c7a9b4a33d3bdb3988d6229e50"
 dependencies = [
  "bit-set",
  "bit-vec",
- "bitflags",
+ "bitflags 2.9.0",
  "lazy_static",
  "num-traits",
  "rand",
@@ -4533,9 +4808,15 @@ version = "11.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "c6df7ab838ed27997ba19a4664507e6f82b41fe6e20be42929332156e5e85146"
 dependencies = [
- "bitflags",
+ "bitflags 2.9.0",
 ]
 
+[[package]]
+name = "rawpointer"
+version = "0.2.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "60a357793950651c4ed0f3f52338f53b2f809f32d83a07f72909fa13e4c6c1e3"
+
 [[package]]
 name = "rayon"
 version = "1.10.0"
@@ -4556,6 +4837,15 @@ dependencies = [
  "crossbeam-utils",
 ]
 
+[[package]]
+name = "redox_syscall"
+version = "0.5.10"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0b8c0c260b63a8219631167be35e6a988e9554dbd323f8bd08439c8ed1302bd1"
+dependencies = [
+ "bitflags 2.9.0",
+]
+
 [[package]]
 name = "redox_users"
 version = "0.4.6"
@@ -5405,7 +5695,7 @@ dependencies = [
  "alloy-eip7702 0.4.2",
  "alloy-primitives",
  "auto_impl",
- "bitflags",
+ "bitflags 2.9.0",
  "bitvec",
  "c-kzg",
  "cfg-if",
@@ -5424,7 +5714,7 @@ dependencies = [
  "alloy-eip7702 0.5.1",
  "alloy-primitives",
  "auto_impl",
- "bitflags",
+ "bitflags 2.9.0",
  "bitvec",
  "c-kzg",
  "cfg-if",
@@ -5444,7 +5734,7 @@ dependencies = [
  "alloy-eip7702 0.5.1",
  "alloy-primitives",
  "auto_impl",
- "bitflags",
+ "bitflags 2.9.0",
  "bitvec",
  "c-kzg",
  "cfg-if",
@@ -5619,7 +5909,7 @@ version = "0.38.44"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "fdb5bc1ae2baa591800df16c9ca78619bf65c0488b41b96ccec5d11220d8c154"
 dependencies = [
- "bitflags",
+ "bitflags 2.9.0",
  "errno",
  "libc",
  "linux-raw-sys",
@@ -5743,6 +6033,12 @@ dependencies = [
  "thiserror 1.0.69",
 ]
 
+[[package]]
+name = "scopeguard"
+version = "1.2.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "94143f37725109f92c262ed2cf5e59bce7498c01bcc1502d7b9afe439a4e9f49"
+
 [[package]]
 name = "scroll-alloy-consensus"
 version = "1.1.5"
@@ -5899,6 +6195,7 @@ dependencies = [
  "openvm-native-recursion",
  "openvm-native-transpiler",
  "openvm-sdk",
+ "openvm-stark-sdk",
  "openvm-transpiler",
  "sbv",
  "scroll-zkvm-circuit-input-types",
@@ -6279,6 +6576,15 @@ dependencies = [
  "der",
 ]
 
+[[package]]
+name = "sppark"
+version = "0.1.8"
+source = "git+https://github.com/scroll-tech/sppark.git?branch=sp1-v3.0#07fe0de43a745a395dd718c735cdfe0cfc457b0a"
+dependencies = [
+ "cc",
+ "which",
+]
+
 [[package]]
 name = "stability"
 version = "0.2.1"
@@ -6929,6 +7235,18 @@ version = "0.1.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "ba73ea9cf16a25df0c8caa16c51acb937d5712a8429db78a3ee29d5dcacd3a65"
 
+[[package]]
+name = "vek"
+version = "0.15.10"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8085882662f9bc47fc8b0cdafa5e19df8f592f650c02b9083da8d45ac9eebd17"
+dependencies = [
+ "approx",
+ "num-integer",
+ "num-traits",
+ "rustc_version 0.4.1",
+]
+
 [[package]]
 name = "version_check"
 version = "0.9.5"
@@ -7041,6 +7359,18 @@ dependencies = [
  "wasm-bindgen",
 ]
 
+[[package]]
+name = "which"
+version = "4.4.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "87ba24419a2078cd2b0f2ede2691b6c66d8e47836da3b6db8265ebad47afbfc7"
+dependencies = [
+ "either",
+ "home",
+ "once_cell",
+ "rustix",
+]
+
 [[package]]
 name = "winapi"
 version = "0.3.9"
@@ -7285,7 +7615,7 @@ version = "0.33.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "3268f3d866458b787f390cf61f4bbb563b922d091359f9608842999eaee3943c"
 dependencies = [
- "bitflags",
+ "bitflags 2.9.0",
 ]
 
 [[package]]
diff --git a/Cargo.toml b/Cargo.toml
index c599eef3..7338ad16 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -39,7 +39,7 @@ openvm-rv32im-transpiler = { git = "https://github.com/openvm-org/openvm.git", r
 openvm-transpiler = { git = "https://github.com/openvm-org/openvm.git", rev = "f1b4844", default-features = false }
 openvm-sdk = { git = "https://github.com/openvm-org/openvm.git", rev = "f1b4844", default-features = false, features = ["parallel", "bench-metrics"] }
 
-openvm-stark-sdk = { git = "https://github.com/openvm-org/stark-backend.git", rev = "bc364134b8315c27bfd29c6e77ac79fe77090137", default-features = false }
+openvm-stark-sdk = { git = "https://github.com/openvm-org/stark-backend.git", rev = "bc364134b8315c27bfd29c6e77ac79fe77090137" }
 
 bitcode = { version = "0.6.3", default-features = false, features = ["serde", "derive"] }
 derivative = "2.2.0"
@@ -76,6 +76,37 @@ alloy-primitives = { git = "https://github.com/scroll-tech/alloy-core", branch =
 ruint = { git = "https://github.com/scroll-tech/uint.git", branch = "v1.12.3" }
 tiny-keccak = { git = "https://github.com/scroll-tech/tiny-keccak", branch = "scroll-patch-v2.0.2-openvm-v1.0.0-rc.1" }
 
+[patch."https://github.com/openvm-org/stark-backend.git"]
+openvm-stark-backend = { git = "ssh://git@github.com/scroll-tech/openvm-stark-gpu.git", branch = "main", features = ["gpu"] }
+openvm-stark-sdk = { git = "ssh://git@github.com/scroll-tech/openvm-stark-gpu.git", branch = "main", features = ["gpu"] }
+
+[patch."https://github.com/Plonky3/Plonky3.git"]
+p3-air = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" }
+p3-field = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" }
+p3-commit = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" }
+p3-matrix = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" }
+p3-baby-bear = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", features = [
+    "nightly-features",
+], branch = "openvm-v2" }
+p3-util = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" }
+p3-challenger = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" }
+p3-dft = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" }
+p3-fri = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" }
+p3-goldilocks = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" }
+p3-keccak = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" }
+p3-keccak-air = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" }
+p3-blake3 = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" }
+p3-mds = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" }
+p3-merkle-tree = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" }
+p3-monty-31 = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" }
+p3-poseidon = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" }
+p3-poseidon2 = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" }
+p3-poseidon2-air = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" }
+p3-symmetric = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" }
+p3-uni-stark = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" }
+p3-maybe-rayon = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" } # the "parallel" feature is NOT on by default to allow single-threaded benchmarking
+p3-bn254-fr = { git = "ssh://git@github.com/scroll-tech/plonky3-gpu.git", branch = "openvm-v2" }
+
 [profile.release]
 codegen-units = 1
 panic = "abort"
diff --git a/crates/integration/Cargo.toml b/crates/integration/Cargo.toml
index f9bc32de..b2a75bac 100644
--- a/crates/integration/Cargo.toml
+++ b/crates/integration/Cargo.toml
@@ -19,6 +19,7 @@ openvm-native-compiler = { workspace = true, default-features = false }
 openvm-native-recursion = { workspace = true, default-features = false }
 openvm-native-transpiler = { workspace = true, default-features = false }
 openvm-transpiler = { workspace = true, default-features = false }
+openvm-stark-sdk.workspace = true
 
 alloy-primitives.workspace = true
 tiny-keccak.workspace = true
@@ -41,3 +42,4 @@ glob = "0.3"
 default = ["scroll"]
 limit-logs = []
 scroll = ["sbv/scroll"]
+bench-metrics = ["scroll-zkvm-prover/bench-metrics", "openvm-sdk/bench-metrics", "openvm-circuit/bench-metrics"]
diff --git a/crates/integration/tests/batch_circuit.rs b/crates/integration/tests/batch_circuit.rs
index dd150564..8af8115c 100644
--- a/crates/integration/tests/batch_circuit.rs
+++ b/crates/integration/tests/batch_circuit.rs
@@ -1,3 +1,4 @@
+use openvm_stark_sdk::bench::run_with_metric_collection;
 use scroll_zkvm_integration::{
     ProverTester, prove_verify_multi, prove_verify_single,
     testers::{
@@ -46,9 +47,11 @@ fn setup_prove_verify_single() -> eyre::Result<()> {
     BatchProverTester::setup()?;
 
     let task = load_recent_chunk_proofs()?;
-    prove_verify_single::<BatchProverTester>(Some(task))?;
+    run_with_metric_collection("OUTPUT_PATH", || -> eyre::Result<()> {
+        prove_verify_single::<BatchProverTester>(Some(task))?;
 
-    Ok(())
+        Ok(())
+    })
 }
 
 #[test]
diff --git a/crates/integration/tests/bundle_circuit.rs b/crates/integration/tests/bundle_circuit.rs
index 869091f5..624067f2 100644
--- a/crates/integration/tests/bundle_circuit.rs
+++ b/crates/integration/tests/bundle_circuit.rs
@@ -1,3 +1,4 @@
+use openvm_stark_sdk::bench::run_with_metric_collection;
 use scroll_zkvm_integration::{
     ProverTester, prove_verify_multi, prove_verify_single_evm,
     testers::{
@@ -25,9 +26,11 @@ fn setup_prove_verify() -> eyre::Result<()> {
     BundleProverTester::setup()?;
 
     let task = load_recent_batch_proofs()?;
-    prove_verify_single_evm::<BundleProverTester>(Some(task))?;
+    run_with_metric_collection("OUTPUT_PATH", || -> eyre::Result<()> {
+        prove_verify_single_evm::<BundleProverTester>(Some(task))?;
 
-    Ok(())
+        Ok(())
+    })
 }
 
 #[test]
diff --git a/crates/integration/tests/chunk_circuit.rs b/crates/integration/tests/chunk_circuit.rs
index 467f8183..79ec28e2 100644
--- a/crates/integration/tests/chunk_circuit.rs
+++ b/crates/integration/tests/chunk_circuit.rs
@@ -1,3 +1,4 @@
+use openvm_stark_sdk::bench::run_with_metric_collection;
 use scroll_zkvm_integration::{
     ProverTester, prove_verify_multi, prove_verify_single,
     testers::chunk::{ChunkProverTester, MultiChunkProverTester},
@@ -20,9 +21,11 @@ fn test_execute() -> eyre::Result<()> {
 fn setup_prove_verify_single() -> eyre::Result<()> {
     ChunkProverTester::setup()?;
 
-    prove_verify_single::<ChunkProverTester>(None)?;
+    run_with_metric_collection("OUTPUT_PATH", || -> eyre::Result<()> {
+        prove_verify_single::<ChunkProverTester>(None)?;
 
-    Ok(())
+        Ok(())
+    })
 }
 
 #[test]
diff --git a/crates/prover/Cargo.toml b/crates/prover/Cargo.toml
index 56693f43..209d8653 100644
--- a/crates/prover/Cargo.toml
+++ b/crates/prover/Cargo.toml
@@ -41,3 +41,4 @@ eyre = "0.6"
 [features]
 default = ["scroll"]
 scroll = ["sbv/scroll"]
+bench-metrics = ["openvm-sdk/bench-metrics", "openvm-circuit/bench-metrics"]
diff --git a/crates/prover/src/prover/chunk.rs b/crates/prover/src/prover/chunk.rs
index 37a6c786..9d457c33 100644
--- a/crates/prover/src/prover/chunk.rs
+++ b/crates/prover/src/prover/chunk.rs
@@ -50,6 +50,7 @@ impl ProverType for ChunkProverType {
         path_app_config: P,
     ) -> Result<openvm_sdk::config::AppConfig<openvm_sdk::config::SdkVmConfig>, Error> {
         let mut config = read_app_config(path_app_config)?;
+        assert_eq!(config.app_fri_params.fri_params.log_blowup, 1);
         config.app_vm_config.system.config = config
             .app_vm_config
             .system