From WikiChip
Difference between revisions of "intel/microarchitectures/broadwell (client)"
Line 66: | Line 66: | ||
| core name = Broadwell Y | | core name = Broadwell Y | ||
| core name 2 = Broadwell U | | core name 2 = Broadwell U | ||
− | | core name | + | | core name 3 = Broadwell H |
− | | core name | + | | core name 4 = Broadwell DT |
− | | core name | + | | core name 5 = Broadwell EP |
− | | core name | + | | core name 6 = Broadwell EX |
− | | core name | + | | core name 7 = Broadwell E |
| succession = Yes | | succession = Yes |
Revision as of 21:47, 12 April 2016
Edit Values | |
Broadwell µarch | |
General Info |
Broadwell is Intel's microarchitecture based on the 14 nm process for mobile, desktops, and servers. Introduced in early 2015, Broadwell is a process shrink of Haswell which introduced several enhancements.
Codenames
Core | Target |
---|---|
Broadwell Y | Core M family, SoC for Smartphones, 2-in-1s Tablets, and notebooks |
Broadwell U | Core ultrabooks |
Broadwell H | IoT (QM87, HM86/HM87 Chipsets), All-in-ones |
Broadwell DT | Unlocked desktop MPUs |
Broadwell EP | Xeon E5, Dual-Processor platform |
Broadwell EX | Xeon E5, Multi-Processor platform, QPI |
Broadwell E | High-End Desktops (HEDT) |
Architecture
Broadwell is for the most part identical to Haswell with server enhancements.
Key changes from Haswell
- ~5% IPC improvement
- FP multiplication instructions has reduced latency (3 cycles, down from 5)
- Affects AVX, SSE, and FP instructions
- CLMUL instructions are now a single μop, improving latency and throughput
- The second-level TLB (STLB)
- Table was enlarged (1,536 entries, up from 1024)
- 1GB page mode (16 entries, 4-ways set associative)
- Larger out-of-order scheduler
- Faster store-to-load forwarding
- Address prediction for branches and returns was improved
- Improved cryptography acceleration instructions
Core features maintained a 2:1 ratio of performance:power.
Graphics
- 50% higher sampler throughput
- Improvements for increased geometry, Z, Pixel Fill
- Direct X 11.2, OpenGL 4.3
- OpenCL 1.2 and 2.0 (with Shared Virtual Memory)
- Up to 24 EUs (20% addition, up from 20 in Haswell), 48 EUs on Iris Pro Graphics