QUAC 100 Release Notes

v1.2.0 January 2026 FIPS 140-3 (IUT — atsec)

New Features #

QUAC 100 firmware v1.2.0 introduces significant performance improvements and expanded algorithm support over the previous v1.1.x release series:

ML-KEM Performance Optimization

The Radix-32 NTT engine has been further optimized to deliver 1.4 million ML-KEM-768 encapsulation operations per second, representing a 12% improvement over v1.1.3. Latency for single operations has been reduced to sub-700 nanosecond end-to-end for the complete encapsulation cycle.

ML-DSA Batch Signing

New batch signing mode enables up to 850,000 ML-DSA-65 signature operations per second when processing queued requests. The job dispatcher now supports priority-weighted fair queuing across multiple cryptographic contexts.

QRNG Throughput Enhancement

Quantum random number generation conditioned output has been increased to >800 Mbps (up from 400 Mbps in v1.1.3) through improvements to the dual free-running ring oscillator entropy collection pipeline. Raw pre-conditioning bitrate is approximately 1.5 Gbps. Real-time NIST SP 800-90B compliance monitoring is now integrated into the conditioning path.

See also: For complete performance benchmarks, refer to the Hardware Architecture documentation.

Resolved Issues #

The following issues have been resolved in firmware v1.2.0:

Table 1 — Resolved Issues in v1.2.0
Issue IDComponentDescription
QUAC-1847NTT EngineFixed occasional coefficient overflow during ML-KEM-1024 decapsulation under sustained high-throughput load
QUAC-1823PCIe InterfaceResolved DMA completion timeout on certain AMD EPYC 9004 series platforms with IOMMU enabled
QUAC-1801HSM ModuleCorrected key slot deallocation race condition during concurrent multi-tenant key lifecycle operations
QUAC-1789QRNGFixed entropy pool starvation during sustained high-throughput consumption with multiple simultaneous consumers
QUAC-1776Power MgmtResolved incorrect C-state transition causing 200μs wake latency spike in dynamic power scaling mode

Known Issues #

Table 2 — Known Issues
Issue IDComponentDescriptionWorkaround
QUAC-1862SR-IOVVF allocation may fail when requesting more than 200 concurrent virtual functions on VMware ESXi 8.0u2Limit to 192 VFs or use KVM
QUAC-1855FirmwareSecure boot attestation chain verification adds 50ms to cold boot sequenceNone; fix targeted for v1.2.1

Compatibility Matrix #

Table 3 — Platform Compatibility
ComponentSupported Versions
Linux Kernel5.15+, 6.x (Ubuntu 22.04/24.04, RHEL 8.8+/9.x, SLES 15 SP5+)
WindowsWindows Server 2022, Windows 11 23H2+
PCIeGen5 x8x8 (or Gen4 x16; backward compatible Gen3 x16 at reduced throughput)
QuantaCore SDKv2.1.0+ required for v1.2.0 firmware features
HypervisorsVMware ESXi 7.0u3+, KVM/QEMU 7.0+, Xen 4.17+
CPU PlatformsIntel Xeon Scalable 4th/5th Gen, AMD EPYC 7003/9004, ARM Neoverse N2+

Downloads #

Table 4 — Firmware & Driver Downloads
PackageVersionPlatform
QUAC 100 Firmwarev1.2.0All platforms
Linux Driverv2.1.0Linux x86_64, aarch64
Windows Driverv2.1.0Windows Server 2022, Win11
QuantaCore SDKv2.1.0Linux, Windows, macOS (dev only)
Important: Always update the QuantaCore SDK to the matching version when upgrading firmware. Mismatched firmware/SDK versions may result in degraded performance or feature incompatibility.

Previous Releases #

v1.1.3 — November 2025

Bug fix release addressing PCIe link training stability on AMD platforms and improving ML-DSA verification throughput by 8%. Added preliminary SLH-DSA-128s support in evaluation mode.

v1.1.0 — September 2025

Major feature release introducing SR-IOV support for up to 252 virtual functions, QuantaVirt management plane integration, and initial FIPS 140-3 validation submission artifacts.

v1.0.0 — June 2025

Initial production release with ML-KEM-768/1024, ML-DSA-44/65/87, hardware QRNG, and integrated HSM functionality.