Skip to content
‹ All benchmarks

CT × CT matrix multiplication

N = 65,536 · Scale 250· Encrypted matrix × encrypted matrix

ScalingThroughput
Log Y

Workload performance

HoistingOn

Complete workload

64 × 64 · Entry depth 0
NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionNVIDIA GeForce RTX 4060 TiNVIDIA RTX A6000 · 300 WNVIDIA H100 NVL · 400 WAMD Ryzen Threadripper PRO 7975WX 32-Cores
Throughput (tasks/s)08.7517.526.335124816Batch sizeNVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 28.3NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 26.3NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 22.9NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 21NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 19.7NVIDIA GeForce RTX 4060 Ti · 4.36NVIDIA GeForce RTX 4060 Ti · 3.9NVIDIA GeForce RTX 4060 Ti · 3.7NVIDIA RTX A6000 · 300 W · 7.33NVIDIA RTX A6000 · 300 W · 7.45NVIDIA RTX A6000 · 300 W · 7.53NVIDIA RTX A6000 · 300 W · 7.58NVIDIA RTX A6000 · 300 W · 7.6NVIDIA H100 NVL · 400 W · 21.3NVIDIA H100 NVL · 400 W · 21.2NVIDIA H100 NVL · 400 W · 21.4NVIDIA H100 NVL · 400 W · 21.6NVIDIA H100 NVL · 400 W · 21.7AMD Ryzen Threadripper PRO 7975WX 32-Cores · 0.918AMD Ryzen Threadripper PRO 7975WX 32-Cores · 0.514AMD Ryzen Threadripper PRO 7975WX 32-Cores · 0.494AMD Ryzen Threadripper PRO 7975WX 32-Cores · 0.479AMD Ryzen Threadripper PRO 7975WX 32-Cores · 0.494

Operations

6 operations

Rotate (1 offset)

Entry depth 0 · Shift +1
NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionNVIDIA GeForce RTX 4060 TiNVIDIA RTX A6000 · 300 WNVIDIA H100 NVL · 400 WAMD Ryzen Threadripper PRO 7975WX 32-Cores
Throughput (tasks/s)0225450675900124816Batch sizeNVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 775NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 758NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 661NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 549NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 488NVIDIA GeForce RTX 4060 Ti · 116NVIDIA GeForce RTX 4060 Ti · 103NVIDIA GeForce RTX 4060 Ti · 94.9NVIDIA GeForce RTX 4060 Ti · 94.8NVIDIA GeForce RTX 4060 Ti · 94.9NVIDIA RTX A6000 · 300 W · 173NVIDIA RTX A6000 · 300 W · 176NVIDIA RTX A6000 · 300 W · 179NVIDIA RTX A6000 · 300 W · 180NVIDIA RTX A6000 · 300 W · 180NVIDIA H100 NVL · 400 W · 513NVIDIA H100 NVL · 400 W · 533NVIDIA H100 NVL · 400 W · 531NVIDIA H100 NVL · 400 W · 536NVIDIA H100 NVL · 400 W · 507AMD Ryzen Threadripper PRO 7975WX 32-Cores · 24.1AMD Ryzen Threadripper PRO 7975WX 32-Cores · 22.1AMD Ryzen Threadripper PRO 7975WX 32-Cores · 17.2AMD Ryzen Threadripper PRO 7975WX 32-Cores · 17AMD Ryzen Threadripper PRO 7975WX 32-Cores · 16.6

Rotate-many (7 offsets)

HoistingOn
Entry depth 0 · Shifts +1…+7 · one input
NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionNVIDIA GeForce RTX 4060 TiNVIDIA RTX A6000 · 300 WNVIDIA H100 NVL · 400 WAMD Ryzen Threadripper PRO 7975WX 32-Cores
Throughput (tasks/s)062.5125188250124816Batch sizeNVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 203NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 181NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 139NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 127NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 122NVIDIA GeForce RTX 4060 Ti · 24.7NVIDIA GeForce RTX 4060 Ti · 22NVIDIA GeForce RTX 4060 Ti · 21.5NVIDIA GeForce RTX 4060 Ti · 21.3NVIDIA GeForce RTX 4060 Ti · 21.3NVIDIA RTX A6000 · 300 W · 45.7NVIDIA RTX A6000 · 300 W · 46.1NVIDIA RTX A6000 · 300 W · 46.5NVIDIA RTX A6000 · 300 W · 46.8NVIDIA RTX A6000 · 300 W · 46.9NVIDIA H100 NVL · 400 W · 151NVIDIA H100 NVL · 400 W · 151NVIDIA H100 NVL · 400 W · 144NVIDIA H100 NVL · 400 W · 147NVIDIA H100 NVL · 400 W · 147AMD Ryzen Threadripper PRO 7975WX 32-Cores · 6.47AMD Ryzen Threadripper PRO 7975WX 32-Cores · 4.97AMD Ryzen Threadripper PRO 7975WX 32-Cores · 4.59AMD Ryzen Threadripper PRO 7975WX 32-Cores · 4.36AMD Ryzen Threadripper PRO 7975WX 32-Cores · 4.03

Multiply (CT × CT)

Entry depth 0 · Distinct inputs · no relinearization or rescale
NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionNVIDIA GeForce RTX 4060 TiNVIDIA RTX A6000 · 300 WNVIDIA H100 NVL · 400 WAMD Ryzen Threadripper PRO 7975WX 32-Cores
Throughput (tasks/s)08,75017,50026,30035,000124816Batch sizeNVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 29,700NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 15,600NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 13,900NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 14,200NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 14,300NVIDIA GeForce RTX 4060 Ti · 2,200NVIDIA GeForce RTX 4060 Ti · 2,310NVIDIA GeForce RTX 4060 Ti · 2,370NVIDIA GeForce RTX 4060 Ti · 2,400NVIDIA GeForce RTX 4060 Ti · 2,430NVIDIA RTX A6000 · 300 W · 5,940NVIDIA RTX A6000 · 300 W · 6,250NVIDIA RTX A6000 · 300 W · 6,410NVIDIA RTX A6000 · 300 W · 6,490NVIDIA RTX A6000 · 300 W · 6,530NVIDIA H100 NVL · 400 W · 20,000NVIDIA H100 NVL · 400 W · 23,700NVIDIA H100 NVL · 400 W · 24,400NVIDIA H100 NVL · 400 W · 24,600NVIDIA H100 NVL · 400 W · 24,700AMD Ryzen Threadripper PRO 7975WX 32-Cores · 942AMD Ryzen Threadripper PRO 7975WX 32-Cores · 457AMD Ryzen Threadripper PRO 7975WX 32-Cores · 516AMD Ryzen Threadripper PRO 7975WX 32-Cores · 491AMD Ryzen Threadripper PRO 7975WX 32-Cores · 408

Add (CT + CT)

Entry depth 0 · Same depth and scale
NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionNVIDIA GeForce RTX 4060 TiNVIDIA RTX A6000 · 300 WNVIDIA H100 NVL · 400 WAMD Ryzen Threadripper PRO 7975WX 32-Cores
Throughput (tasks/s)012,50025,00037,50050,000124816Batch sizeNVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 40,400NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 19,700NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 16,600NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 16,900NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 17,100NVIDIA GeForce RTX 4060 Ti · 2,620NVIDIA GeForce RTX 4060 Ti · 2,730NVIDIA GeForce RTX 4060 Ti · 2,810NVIDIA GeForce RTX 4060 Ti · 2,920NVIDIA GeForce RTX 4060 Ti · 2,960NVIDIA RTX A6000 · 300 W · 6,890NVIDIA RTX A6000 · 300 W · 7,280NVIDIA RTX A6000 · 300 W · 7,510NVIDIA RTX A6000 · 300 W · 7,630NVIDIA RTX A6000 · 300 W · 7,670NVIDIA H100 NVL · 400 W · 24,800NVIDIA H100 NVL · 400 W · 30,600NVIDIA H100 NVL · 400 W · 34,700NVIDIA H100 NVL · 400 W · 37,200NVIDIA H100 NVL · 400 W · 38,200AMD Ryzen Threadripper PRO 7975WX 32-Cores · 5,690AMD Ryzen Threadripper PRO 7975WX 32-Cores · 3,630AMD Ryzen Threadripper PRO 7975WX 32-Cores · 797AMD Ryzen Threadripper PRO 7975WX 32-Cores · 863AMD Ryzen Threadripper PRO 7975WX 32-Cores · 927

Relinearize

Entry depth 0 · 3 → 2 components · no rescale
NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionNVIDIA GeForce RTX 4060 TiNVIDIA RTX A6000 · 300 WNVIDIA H100 NVL · 400 WAMD Ryzen Threadripper PRO 7975WX 32-Cores
Throughput (tasks/s)0225450675900124816Batch sizeNVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 781NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 770NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 676NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 558NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 498NVIDIA GeForce RTX 4060 Ti · 121NVIDIA GeForce RTX 4060 Ti · 106NVIDIA GeForce RTX 4060 Ti · 97.6NVIDIA GeForce RTX 4060 Ti · 97.4NVIDIA GeForce RTX 4060 Ti · 97.4NVIDIA RTX A6000 · 300 W · 174NVIDIA RTX A6000 · 300 W · 178NVIDIA RTX A6000 · 300 W · 180NVIDIA RTX A6000 · 300 W · 181NVIDIA RTX A6000 · 300 W · 182NVIDIA H100 NVL · 400 W · 519NVIDIA H100 NVL · 400 W · 528NVIDIA H100 NVL · 400 W · 514NVIDIA H100 NVL · 400 W · 516NVIDIA H100 NVL · 400 W · 498AMD Ryzen Threadripper PRO 7975WX 32-Cores · 24.6AMD Ryzen Threadripper PRO 7975WX 32-Cores · 20.7AMD Ryzen Threadripper PRO 7975WX 32-Cores · 18.1AMD Ryzen Threadripper PRO 7975WX 32-Cores · 17.7AMD Ryzen Threadripper PRO 7975WX 32-Cores · 15.8

Rescale (2 components)

Entry depth 0 · One Q depth group
NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionNVIDIA GeForce RTX 4060 TiNVIDIA RTX A6000 · 300 WNVIDIA H100 NVL · 400 WAMD Ryzen Threadripper PRO 7975WX 32-Cores
Throughput (tasks/s)01,5003,0004,5006,000124816Batch sizeNVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 4,790NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 5,190NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 4,840NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 3,530NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 3,300NVIDIA GeForce RTX 4060 Ti · 799NVIDIA GeForce RTX 4060 Ti · 578NVIDIA GeForce RTX 4060 Ti · 586NVIDIA GeForce RTX 4060 Ti · 590NVIDIA GeForce RTX 4060 Ti · 594NVIDIA RTX A6000 · 300 W · 1,070NVIDIA RTX A6000 · 300 W · 1,100NVIDIA RTX A6000 · 300 W · 1,110NVIDIA RTX A6000 · 300 W · 1,110NVIDIA RTX A6000 · 300 W · 1,120NVIDIA H100 NVL · 400 W · 2,900NVIDIA H100 NVL · 400 W · 3,110NVIDIA H100 NVL · 400 W · 3,250NVIDIA H100 NVL · 400 W · 3,170NVIDIA H100 NVL · 400 W · 3,090AMD Ryzen Threadripper PRO 7975WX 32-Cores · 143AMD Ryzen Threadripper PRO 7975WX 32-Cores · 170AMD Ryzen Threadripper PRO 7975WX 32-Cores · 88.5AMD Ryzen Threadripper PRO 7975WX 32-Cores · 76.9AMD Ryzen Threadripper PRO 7975WX 32-Cores · 93.8

Measurement conditions

ConfigurationN = 65,536 · scale 250
ExecutionEager · prepared inputs and keys
PlatformsNVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition / NVIDIA GeForce RTX 4060 Ti / NVIDIA RTX A6000 · 300 W / NVIDIA H100 NVL · 400 W / AMD Ryzen Threadripper PRO 7975WX 32-Cores