Steamroller (microarchitecture)

Steamroller - Family 15h (3rd-gen)

Produced	beginning of 2014
Common manufacturer(s)	AMD
Min. feature size	28 nm SHP[1]
Instruction set	AMD64 (x86-64)
Socket(s)	Socket FM2+ Socket FP3 (µBGA)
Predecessor	Piledriver - Family 15h (2nd-gen)
Successor	Excavator - Family 15h (4th-gen)
Core name(s)	AMD APU[2]

AMD Steamroller Family 15h is a microarchitecture developed by AMD for AMD APUs, which succeeded Piledriver in the beginning of 2014 as the third-generation Bulldozer-based microarchitecture.^[3] Steamroller APUs continue to use two-core modules as their predecessors, while aiming at achieving greater levels of parallelism.

Microarchitecture

Steamroller still features two-core modules found in Bulldozer and Piledriver designs called clustered multi-thread (CMT), meaning that one module is equal to a dual-core processor.^[4] The focus of Steamroller is for greater parallelism.^[5] Improvements center on independent instruction decoders for each core within a module, 25% more of the maximum width dispatches per thread, better instruction schedulers, improved perceptron branch predictor, larger and smarter caches, up to 30% fewer instruction cache misses, branch misprediction rate reduced by 20%, dynamically resizable L2 cache, micro-operations queue,^[6] more internal register resources and improved memory controller. Another improvement over the Piledriver cores were the addition of new CPU instructions, such as HEVC.

AMD estimated that these improvements will increase instructions per cycle (IPC) up to 30% compared to the first-generation Bulldozer core while maintaining Piledriver's high clock rates with decreased power consumption.^[4] The final result was a 9% single-threaded IPC improvement, and 18% multi-threaded IPC improvement over Piledriver.^[7]

Steamroller, the microarchitecture for CPUs, as well as Graphics Core Next, the microarchitecture for GPUs, are paired together in certain product lines, and both support certain features specified in Heterogeneous System Architecture in hardware.

History

In 2011, AMD announced a third-generation Bulldozer-based line of processors for 2013,^[8] with Next Generation Bulldozer as the working title, using the 28 nm manufacturing process.^[9]

On 21 September 2011, leaked AMD slides indicated that this third generation of Bulldozer core was codenamed Steamroller.^[10][11]

In January 2014, Kaveri APUs became available.^[12]

Processors

APU lines

There are two main APU lines announced:

1. Kaveri A-series APU
   • Desktop budget and mainstream markets (FM2+): The Trinity / Richland APU line was replaced in January 2014 by the Kaveri APU line, as the third generation of A10, A8, A6 and A4 series for the desktop market. Currently known new model is a quad-core A10-7850K APU, with a 3.7 GHz core frequency and 4 MB L2 cache, incorporating a 720 MHz GPU with 512 stream processors and over 856 GFLOPS of total processing power.^[13]
     An upcoming product is the A10-7890K. These APUs have two to four enhanced Steamroller B cores, a GCN 1.1 Volcanic Islands (Sea Islands for Godavari / Kaveri-refresh A10-7870K integrated GPU,^[14][15] and two integrated DDR3 memory controllers. Kaveri APUs utilize the new FM2+ socket, thus they are not backwards compatible with the previous generation of FM2 motherboards.^[16]
   • SIP blocks: Unified Video Decoder, Video Coding Engine, TrueAudio^[17]
   • AMD Eyefinity up to 4 monitors,^[18] 4K Ultra HD support, DisplayPort 1.2 Support^[19]
   • Socket FM2+-only, Socket FM2 is not supported,^[20] support for PCIe 3.0
   • Two or four CPU cores based on the Steamroller microarchitecture
   • Three to eight Compute Units (CUs) based on the revised Graphics Core Next (GCN)^[21] microarchitecture; 1 Compute Unit (CU) consists of 64 Unified Shader Processors : 4 Texture Mapping Units (TMUs) : 1 Render Output Unit (ROPs)
   • AMD Heterogeneous System Architecture (HSA) 2.0
   • Dual-channel (2x64-bit) DDR3 memory controller
   • Integrated custom ARM Cortex-A5 co-processor^[22] with TrustZone Security Extensions^[23]
   • Select models support AMD Hybrid Graphics by using a Radeon R7 240 or R7 250 discrete graphics card.^[24]
2. Berlin APU
   • Enterprise and server markets: The Berlin APU will be similar to Kaveri, featuring four Steamroller cores, up to 512 stream processors, and support for ECC memory.^[25]

FX lines

In November 2013, AMD confirmed it will not update the FX series in 2014, neither its current Socket AM3+ version, nor will it receive a Steamroller version with a new socket.^[26][27] An unconfirmed article, citing anonymous sources with knowledge of AMD's internal roadmap extending into 2015, also did not mention plans for a new FX processor.^[28]

Server lines

AMD's server roadmaps for 2014 show:^[29][30]

• Berlin APU - quad-core x86 Steamroller architecture (as described above) for 1 Processor (1P) compute and media clusters
  • focused on cloud and media streaming
• Berlin CPU - quad-core x86 Steamroller architecture for 1P web and enterprise services clusters
  • focused on big data processing and HPC
• Seattle CPU - 4/8 core AArch64 Cortex-A57 architecture (Opteron A1100) for 1P web and enterprise services clusters ^[31]
  • low-end applications with power-saving options
• Warsaw CPU - up to 16 core x86 Piledriver (2nd gen Bulldozer) architecture (Opteron 6338P and 6370P) for 2P/4P servers ^[32]
  • no plans for Steamroller (3rd gen Bulldozer) or Excavator (4th gen Bulldozer) architectures on high-end multi-processor (2P/4P) platforms

However, plans for Steamroller Opteron products were cancelled, likely due to the poor energy efficiency achieved in this generation of the Bulldozer architecture. Energy efficiency was greatly increased in the following generation, (Excavator), which exceeded Jaguar in performance per watt, and approximately doubled performance/watt over Steamroller (for example 20.74 pt/W vs 10.85 pt/W when comparing similar mobile APUs using rough arbitrary metrics). ^{[33] [34]}

References

1. "Page 2 - AMD Kaveri A10-7850K and A8-7600 review: Was it worth the wait for the first true heterogeneous chip?". ExtremeTech. Retrieved 2014-02-19.
2. "AMD Vows Not to Drop Microprocessor Sockets in Next Two Years". X-bit labs. 2012-12-05. Retrieved 2013-09-29.
3. "AMD Kaveri Review: A8-7600 and A10-7850K Tested". Anandtech.com. 2014-01-14. Retrieved 2014-02-08.
4. "AMD: We Are On Track With Steamroller Micro-Architecture in 2013". X-bit labs. 2013-03-31. Retrieved 2013-09-29.
5. Su, Lisa (2012-02-02). "Consumerization, Cloud, Convergence." (PDF). AMD 2012 Financial Analyst Day. Sunnyvale, California: Advanced Micro Devices. p. 26. Retrieved 2012-02-04.
6. Anand Lal Shimpi (2012-08-28). "AMD's Steamroller Detailed: 3rd Generation Bulldozer Core". Retrieved 2013-11-16.
7. Miller, Michael J. (2014-02-14). "Ivytown, Steamroller, 14 and 16nm Process Highlight ISSCC". Forwardthinking.pcmag.com. Retrieved 2014-02-19.
8. Anton Shilov (2010-11-09). "AMD Plans to Release Twenty-Core Microprocessor in 2012". X-bit labs. Retrieved 2012-01-23.
9. "2012 Financial Analyst Day". 2012-02-02. Retrieved 2013-09-29.
10. "Hosszútávú mobil útiterv szivárgott ki az AMD-től - PROHARDVER! Processzor hír". Prohardver.hu. 2011-09-21. Retrieved 2012-01-23.
11. "Nuove roadmap AMD sulle future APU in programma nel 2012 e nel 2013 per il mercato mobile". 2011-09-21. Retrieved 2012-01-23.
12. Joel Hruska (2014-01-14). "AMD Kaveri A10-7850K and A8-7600 review: Was it worth the wait for the first true heterogeneous chip?". extremetech.com. Retrieved 2014-01-17.
13. "AMD Unleashes More Details About Kaveri: HSA, TrueAudio, Mantle".
14. "AMD Unveils Innovative New APUs and SoCs that Give Consumers a More Exciting and Immersive Experience". Retrieved 2013-09-29.
15. Joel Hruska (2013-11-11). "AMD Confirms Kaveri Integrated Graphics Has 512 GPU Cores, Runs Battle Field 4 at APU13". hothardware.com. Retrieved 2013-12-27.
16. "AMD's Next-Gen "Kaveri" APUs Will Require New Mainboards". X-bit labs. May 30, 2013. Retrieved September 29, 2013.
17. "A technical look at AMD's Kaveri architecture". SemiAccurate. 2014-01-15.
18. "Multi-monitor: Civilization V on A10-7850K "Kaveri"".
19. "AMD A8-7600 Kaveri APU review - The Embedded GPU - HSA & hUMA". 2014-01-14.
20. "AMD's Next-Gen "Kaveri" APUs Will Require New Mainboards.". 2013-05-30. Retrieved 06-09-2013. Check date values in: |access-date= (help)
21. "AMD Kaveri APU Architecture Detailed".
22. "AMD to add ARM processors to boost chip security". June 14, 2012. Retrieved September 3, 2013.
23. "AMD and ARM Fusion redefine beyond x86". Retrieved 2013-11-10.
24. "AMD A10-7850K Graphics Performance". Retrieved April 2, 2014.
25. "AMD Berlin Server APU Provides Glimpse At Upcoming Kaveri APU With 4 Steamroller Cores and 512 GCN SPs". 2013-06-19. Retrieved 2013-09-29.
26. Anton Shilov (2013-11-13). "AMD Cans Plans to Introduce Next-Gen FX Microprocessors Next Year". xbitlabs.com.
27. Josh Walrath (2013-09-04). "AMD's Processor Shift: The Future Really is Fusion". Retrieved 2013-09-29.
28. "AMD updates product roadmap for 2014 and 2015". 2013-08-26. Retrieved 2013-09-29.
29. "Berlin, Warsaw are the future of AMD's x86 server lineup". The Tech Report. 2013-06-18. Retrieved 2013-09-29.
30. Mujtaba, Hassan (December 26, 2013). "AMD Opteron Roadmap Reveals Next Generation Toronto and Carrizo APU Details". WCCF Tech. Retrieved January 15, 2015.
31. Mendoza, Menchie (August 13, 2014). "AMD unwraps 64-bit ARM 'Seattle' server chip". Tech Times. Retrieved January 15, 2015.
32. Gasior, Geoff (January 22, 2014). "16-core Warsaw CPUs added to Opteron lineup". The Tech Report. Retrieved January 15, 2015.
33. "Opteron X2150 vs A10 8700P". cpuboss.com. Retrieved November 15, 2016.
34. "AMD A10 8700P vs 7300". cpuboss.com. Retrieved November 15, 2016.