(no commit message)
[libreriscv.git] / 3d_gpu.mdwn
1 # RISC-V 3D GPU / CPU / VPU
2
3 Note: this is a **hybrid** CPU, VPU and GPU. It is not, as many news articles
4 are implying, a "dedicated exclusive GPU". The option exists to **create**
5 a stand-alone GPU product (contact us if this is a product that you want).
6 Our primary goal is to design a **complete** all-in-one processor
7 (System-on-a-Chip) that happens to include libre-licensed VPU and GPU
8 accelerated instructions as part of the actual - main - CPU itself.
9
10 We seek investors, sponsors (whose contributions thanks to NLNet may be
11 tax-deductible), engineers and potential customers, who are
12 interested, as a first product, in the creation and use of an entirely
13 libre low-power mobile class system-on-a-chip. Comparative benchmark
14 performance, pincount and price is the Allwinner A64, except that the
15 power budget target is 2.5 watts in a 16x16mm 320 to 360 pin 0.8mm
16 FBGA package. Instead of single-issue higher clock rate, the design is
17 multi-issue, aiming for around 800mhz.
18
19 The lower pincount, lower power, and higher BGA pitch is all to reduce
20 the cost of product development when it comes to PCB design and layout:
21
22 * Above 4 watts requires metal packages, greater attention to thermal
23 management in the PCB design and layout, and much pricier PMICs.
24 * 0.6mm pitch BGA and below requires much more expensive PCB manufacturing
25 equipment and more costly PCBA techniques.
26 * Above 600 pins begins to reduce production yields as well as increase
27 the cost of testing and packaging.
28
29 We can look at larger higher-power ASICs either later or, if funding
30 is made available, immediately.
31
32 Recent applications to NLNet (Oct 2019) are for a test chip in 180nm,
33 64 bit, single core dual issue, around 300 to 350mhz. This will provide
34 the confidence to go to higher geometries, as well as be a commercially
35 viable embedded product in its own right.
36
37 # Business Objectives
38
39 See [[3d_gpu/business_objectives]]
40
41 * the project shall be a hybrid CPU-GPU-VPU
42 * the project shall be commercial and mass-volume (100 million units
43 and above)
44 * the project shall be entirely transparent so that end-users will be
45 able to trust it
46 * the source code shall be available at all times for all components
47 for BUSINESS reasons, making development and use of SDKs dead simple
48 and aiding and assisting developers AND BUSINESSES in debugging and thus
49 hugely saving them money.
50
51 Reasoning:
52
53 * If the processor is not a hybrid CPU-GPU-VPU, the
54 complexity involved in developing a split shared-memory CPU-GPU both
55 at a hardware and a software level will be so costly it will jeapordise
56 the project.
57 * The project is commercial and mass-volume because there are plenty
58 of academic designs (none of them reaching production where people
59 may benefit), and "Open" designs, created by the Open Hardware
60 Community, sadly due to the high cost of producing ASICs, tend to be
61 focussed on markets that would have been great about twenty to thirty
62 years ago.
63 * Transparency is a key business objective. It is a Unique Selling Point
64 that the processor is developed in a fashion that, should it be
65 independently audited, no opportunity for spying back-door co-processors
66 will be found to have "made their way surreptitiously - or overtly -
67 into the design". Yes, GCHQ: I know about the conversation you had
68 with nCipher (and, to their everlasting credit, that they told you
69 to take a hike)
70
71 # Links:
72
73 * [[shakti/m_class/libre_3d_gpu]]
74 * [[discussion]]
75 * [[resources]]
76 * [[overview]]
77 * [[3d_gpu/funding]]
78 * [[3d_gpu/architecture]]
79 * Founding [[charter]]
80 * Mailing list <http://lists.libre-riscv.org/pipermail/libre-riscv-dev/>
81 * Crowdsupply page <https://www.crowdsupply.com/libre-risc-v/m-class>
82 * Wiki <https://libre-riscv.org>
83 * Git repositories <https://git.libre-riscv.org>
84 * Bugtracker <http://bugs.libre-riscv.org>
85 * Kazan Vulkan Driver (including 3D engine) <https://salsa.debian.org/Kazan-team/kazan>
86 * [NLNet 2019 Milestones](http://bugs.libre-riscv.org/buglist.cgi?columnlist=assigned_to%2Cbug_status%2Cresolution%2Cshort_desc%2Ccf_budget&f1=cf_nlnet_milestone&o1=equals&query_format=advanced&resolution=---&v1=NLnet.2019.02)
87 * NLNet Project Page <https://nlnet.nl/project/Libre-RISCV/>
88 * [[nlnet_proposals]]
89
90 # Progress:
91
92 * Dec 2019: Second round NLNet questions answered. External Review completed. 6 NLNet proposals accepted (EUR 200,000+)
93 * Nov 2019: Alternative FP library to Berkeley softfloat developed. NLNet first round questions answered.
94 * Oct 2019: 3D Standards continued. POWER ISA considered. Open 3D Alliance begins. NLNet funding applications submitted.
95 * Sep 2019: 3D Standards continued. Additional NLNet Funding proposals discussed.
96 * Aug 2019: Development of "Transcendentals" (SIN/COS/ATAN2) Specifications
97 * Jul 2019: Sponsorship from Purism received. IEEE754 FP Mul, Add, DIV,
98 FCLASS and FCVT pipelines completed.
99 * Jun 2019: IEEE754 FP Mul, Add, and FSM "DIV" completed.
100 * May 2019: 6600-style scoreboard started
101 * Apr 2019: NLnet funding approved by independent review committee
102 * Mar 2019: NLnet funding application first and second phase passed
103 * Mar 2019: First successful nmigen pipeline milestone achieved with IEEE754 FADD
104 * Feb 2019: Conversion of John Dawson's IEEE754 FPU to nmigen started
105 * Jan 2019: Second version Simple-V preliminary proposal (suited to LLVM)
106 * 2017 - Nov 2018: Simple-V specification preliminary draft completed
107 * Aug 2018 - Nov 2018: spike-sv implementation of draft spec completed
108 * Aug 2018: Kazan Vulkan Driver initiated
109 * Sep 2018: mailing list established
110 * Sep 2018: Crowdsupply pre-launch page up (for updates)
111 * Dec 2018: preliminary floorplan and architecture designed (comp.arch)
112
113 # News Articles
114
115 * <https://www.phoronix.com/forums/forum/hardware/processors-memory/1133806-libre-risc-v-open-source-effort-now-looking-at-power-instead-of-risc-v/page7>
116 * <https://hub.packtpub.com/a-libre-gpu-effort-based-on-risc-v-rust-llvm-and-vulkan-by-the-developer-of-an-earth-friendly-computer/>
117 * <https://riscv.org/2018/10/packt-hub-article-a-libre-gpu-effort-based-on-risc-v-rust-llvm-and-vulkan-by-the-developer-of-an-earth-friendly-computer/>
118 * <https://www.reddit.com/r/RISCV/comments/9jts9t/theres_a_new_libre_gpu_effort_building_on_riscv/>
119 * <https://www.linux.com/blog/2018/11/risc-v-linux-development-full-swing>
120 * <https://www.phoronix.com/scan.php?page=news_item&px=Libre-GPU-RISC-V-Vulkan>
121 * <https://www.heise.de/newsticker/meldung/Mobilprozessor-mit-freier-GPU-Libre-RISC-V-M-Class-geplant-4242802.html>
122 * <https://news.ycombinator.com/item?id=18094734>
123 * <http://www.tuxmachines.org/node/116004>
124 * <https://linuxfr.org/users/martoni/journaux/risc-v-est-pret-pour-le-desktop>
125 * <https://www.reddit.com/r/hardware/comments/9jlby1/theres_a_new_libre_gpu_effort_building_on_riscv/>
126 * <http://www.eevblog.com/forum/crowd-funded-projects/libre-risc-v-m-class-with-open-source-gpu-and-kazan-vulkan-driver/>
127 * <https://www.reddit.com/domain/libre-riscv.org/>
128 * <https://hardware.slashdot.org/comments.pl?sid=13447940&cid=58160868>
129 * <https://www.phoronix.com/forums/forum/hardware/graphics-cards/1080755-libre-risc-v-gpu-aiming-for-2-5-watt-power-draw-continues-being-plotted/page5>
130 * <https://www.phoronix.com/forums/forum/hardware/processors-memory/1070828-more-details-on-the-proposed-simple-v-extension-to-risc-v-for-gpu-workloads>
131 * <https://slashdot.org/submission/9750302/nlnet-funds-development-of-a-libre-risc-v-3d-cpu>
132 * <https://hardware.slashdot.org/story/19/06/02/0153243/nlnet-funds-development-of-a-libre-risc-v-3d-cpu>
133 * <https://www.phoronix.com/forums/forum/hardware/graphics-cards/1104124-libre-risc-v-snags-50k-eur-grant-to-work-on-its-risc-v-3d-gpu-chip/page6>
134 * <https://news.ycombinator.com/item?id=21112341>
135 * <https://www.reddit.com/r/RISCV/comments/db04j3/libreriscv_3d_cpugpu_seeks_grants_for_ambitious/>
136 * <https://hardware.slashdot.org/story/19/09/29/1845252/libre-risc-v-3d-cpugpu-seeks-grants-for-ambitious-expansion>
137 * <https://forums.puri.sm/t/risc-v-m-class-effort-and-purism-donation/6528/15>
138 * <https://www.pro-linux.de/news/1/27527/comm/1/show-all-comments.html>
139
140 # Information Resources and Tutorials
141
142 * <https://github.com/timvideos/litex-buildenv/wiki/LiteX-for-Hardware-Engineers>
143 * <https://jeffrey.co.in/blog/2014/01/d-flip-flop-using-migen/>
144 * <http://lists.libre-riscv.org/pipermail/libre-riscv-dev/2019-March/000705.html>
145 * <https://chisel.eecs.berkeley.edu/api/latest/chisel3/util/DecoupledIO.html>
146 * <http://www.clifford.at/papers/2016/yosys-synth-formal/slides.pdf>
147 * <http://blog.lambdaconcept.com/doku.php?id=migen:tutorial>
148 * <http://chiselapp.com/user/kc5tja/repository/kestrel-3/dir?ci=6c559135a301f321&name=cores/cpu>
149 * <https://chips4makers.io/blog/>
150 * <https://hackaday.io/project/7817-zynqberry>
151 * <https://wiki.f-si.org/index.php/FSiC2019>
152 * <https://github.com/efabless/raven-picorv32> - <https://efabless.com>
153 * <https://efabless.com/design_catalog/default>
154 * <https://toyota-ai.ventures/>
155 * <https://github.com/lambdaconcept/minerva>
156 * <https://en.wikipedia.org/wiki/Liskov_substitution_principle>
157 * <https://en.wikipedia.org/wiki/Principle_of_least_astonishment>
158 * <https://peertube.f-si.org/videos/watch/379ef007-40b7-4a51-ba1a-0db4f48e8b16>
159 * <https://github.com/riscv/riscv-sbi-doc/blob/master/riscv-sbi.md>
160 * <https://mshahrad.github.io/openpiton-asplos16.html>
161 * <https://wiki.f-si.org/index.php/The_Raven_chip:_First-time_silicon_success_with_qflow_and_efabless>
162 * <http://smallcultfollowing.com/babysteps/blog/2019/04/19/aic-adventures-in-consensus/>
163 * <http://www.crnhq.org/12-Skills-Summary.aspx?rw=c>
164 * <http://bugs.libre-riscv.org/buglist.cgi?columnlist=assigned_to%2Cbug_status%2Cresolution%2Cshort_desc%2Ccf_budget&f1=cf_nlnet_milestone&o1=equals&query_format=advanced&resolution=---&v1=NLnet.2019.02>
165 * <https://pdfs.semanticscholar.org/5060/4e9aff0e37089c4ab9a376c3f35761ffe28b.pdf>
166 * <http://www.acsel-lab.com/arithmetic/arith15/papers/ARITH15_Takagi.pdf>
167 * <https://youtu.be/o5Ihqg72T3c>
168 * <http://flopoco.gforge.inria.fr/>
169 * Fundamentals of Modern VLSI Devices <https://groups.google.com/a/groups.riscv.org/d/msg/hw-dev/b4pPvlzBzu0/7hDfxArEAgAJ>
170
171 # Analog Simulation
172
173 * <https://github.com/Isotel/mixedsim>
174 * <http://www.vlsiacademy.org/open-source-cad-tools.html>
175 * <http://ngspice.sourceforge.net/adms.html>
176 * <https://en.wikipedia.org/wiki/Verilog-AMS#Open_Source_Implementations>
177
178 # Evaluations
179
180 *[[openpower]]