throbber
sae
`Certification of Cranslation
`
`COUNTY OF SUFFOLK
`COMMONWEALTH OF MASSACHUSETTS
`
`October 17, 2022
`
`This is to certify that the attached translation is, to the best of my knowledgeand belief, a true
`and accurate translation from «Japanese»into «English» of the attached document:
`
`«Interview with Ken Kutaragi, President of SCEI (3) - “Why | worked with NVIDIA on the PS3”»
`
`Linguistic Systems, Inc. adheres to an |SO-certified quality management system that ensures
`best practices are always followed in the selection of linguists skilled in both the languages and
`subject matters necessary for every translation.
`
`Bunardatta Wee.
`Benardette McEvoy
`ia
`
`Principal Account Manager
`Linguistic Systems, Inc.
`

`
`¢s) Linguistic Systems, Inc.
`
`260 Franklin Street, Suite 230, Boston MA 02110 ¢ Phone 617-528-7400 ¢ Fax 617-528-7490 * www.linguist.com
`Certifications: ISO 9001 * ISO 17100 * ISO 18587 * ISG 27001
`
`Form OPS-QF-008 (rev.1}
`
`AMD1318_0157001
`
`ATI Ex. 2146
`
`IPR2023-00922
`
`Page 1 of 17
`
`ATI Ex. 2146
`IPR2023-00922
`Page 1 of 17
`
`

`

`TAT/2017
`
`Hiroshige Goto's Weekly Foreign News
`
`aX Yx>—
`
`Hiroshige Goto’s Weekly
`Foreign Newsm
`
`Interview with Ken Kutaragi,
`President of SCEI (3)
`"Why I worked with NVIDIA
`on the PS3 |
`
`
`
`¢Why wedidn’t use a Cell-based
`graphics chip
`
`{Q] Last year we were surprised that
`you didn’t use Cell-based architecture
`when you designed the PS3. Why
`didn’t you build a Cell-based GPU?
`
`
`
`Mr. Ken Kutaragi
`
`{Kutaragi] You can use Cell’s 7 SPEs (Synergistic Processor
`Elements) for graphics. Actually, a few of our models at E3
`didn’t have graphics processors and Cell alone was responsible
`for all the graphics, including rendering. The thing is, using
`Cell for somethinglike this is a bit of a waste. Cell has better
`uses.
`
`Wehad the ideato install two Cells (one to be used mainly
`for graphics), but we decided against it because we thought that
`the importance of using Cell as a computer and its function as a
`shader were too different. Shaders are their own thing, and we
`want them to be architecture that is thoroughly specialized (in
`graphics) that can also be multi-functional. Although,it is also
`possible, for example, to do display mapping (using SPE).
`
`So far (real-time 3D graphics) have madeit look like that is
`happening, butit’s different within the actual (3D graphics)
`space. Still, image resolution has been good so far. Even today,
`most
`
`http://pe watch impress co.jp/docs/2005/0613/kaigail89 htm
`
`V8
`
`AMD1318_0157002
`
`ATI Ex. 2146
`
`IPR2023-00922
`
`Page 2 of 17
`
`ATI Ex. 2146
`IPR2023-00922
`Page 2 of 17
`
`

`

`TAT/2017
`
`Hiroshige Goto's Weekly Foreign News
`
`games coming out on the Xbox 360are still using that kind
`of 3D graphics.
`
`But we wantto do it properly (where 3D metamorphosis is
`reflected in the 3D space). To achievethis, the idea is to share
`data (between the CPU and GPU)as muchas possible. That's
`why weused this architecture. Essentially, the GPU and Cell
`floating-point units should be identical in everything from
`precision to rounding errors. This time, they are muchcloser,
`almost the same. So, we can use (the results of arithmetic) in
`both directions.
`
`«Stopping the use ofDRAM for full HDTV
`
`[Q] I was expecting you to use eDRAM (embedded DRAM)
`for the graphics memory, but I understood why you didn’t use
`eDRAMwhenI heard about the two HDTV screens.
`
`{Kutaragi] Even ifthe GPU doesn't have graphics memory,
`it can be supported by Redwood (a high-speed interface
`connecting Cell and RSX) and YDRAM (a code namefor
`XDRDRAM). Because YDRAM hasunified (memory).
`
`Even so, there is the problem whetherit’s okay to waste
`(bandwidth and cycle time) by accessing memory in distant
`locations when doing (graphics) routine processing and shader
`(computing). There's no need to use (memory bandwidth) on
`Cell for routine tasks. Shaders also need to use memory
`because they perform a tremendous volumeofarithmetic.
`Especially if you want to use two or more screens of 2k x 1k
`(1,920 x 1,080 dots) in progressive display on a full HDTV,
`you will need a large amount of VRAM.
`
`In this scenario eDRAMisn't an option. Using eDRAMis
`good for the PS2. But this timeit’s not enough for even two
`screens. Suppose that an amount of eDRAM (capable of
`supporting a HDTV)is placed in a microchip of 200 or 300
`square millimeters. This would greatly reduce the amount of
`logic (that can be put on the chip dueto the size of the eDRAM)
`and reduce the number of
`
`http://pe watch impress co.jp/docs/2005/0613/kaigail89 htm
`
`2/8
`
`AMD1318_0157003
`
`ATI Ex. 2146
`
`IPR2023-00922
`
`Page 3 of 17
`
`ATI Ex. 2146
`IPR2023-00922
`Page 3 of 17
`
`

`

`TAT/2017
`
`Hiroshige Goto's Weekly Foreign News
`
`shaders. It’s better to use the logic fully and put a lot of shaders
`on it.
`
`¢The shared vision with NVIDIAofthe ideal processor
`
`[{Q]So why did you work with NVIDIAinstead of other GPU
`vendors?
`
`{Kutaragi] Up until now we've made the graphics for Sony
`Computer Entertainment ourselves, together with Toshiba. We
`wentall out and even made our own process technology. But
`now wehave teamed up with NVIDIA to make the computer
`itself.
`
`NVIDIAare going the whole nine yards with PC graphics,
`even doing whatIntel's processors have been doing with
`programmable shaders. NVIDIA is focusing on processor
`features and performance because its developers, including
`NVIDIA's Chief Scientist David B. Kirk, come from various
`computer companies such as SGI. They are the type to pursue
`what they wantto do all the way without worrying aboutthe
`size of the chip. Sometimes they go overboard, but their culture
`is similar to ours.
`
`NVIDIA's approachaligns with our approach in our pursuit
`of working towards a fully programmable processor. I often
`have the opportunity to talk with Jen-Hsun Huang (President
`and CEO of NVIDIA) and David, and oneofthe things that we
`talk about is making the ideal processor. By ideal, naturally we
`mean a processor that surpasses current PCs, or evenall of the
`current processors.
`
`They are movingsteadily in that direction, and they share our
`vision in that sense. They also share our roadmap and are
`influenced by our architecture. We partnered with NVIDIA
`because we understand each other and want to achieve the same
`
`thing.
`
`Anotherfactoris that displays are moving towardsfixed
`pixel displays (such as LCDs). When wehavefixed pixel
`displays, it will be an era where everything including TVs and
`PCs will be combined together. So we want to support
`everything perfectly.
`
`http://pe watch impress co.jp/docs/2005/0613/kaigail89 htm
`
`3/8
`
`AMD1318_0157004
`
`ATI Ex. 2146
`
`IPR2023-00922
`
`Page 4 of 17
`
`ATI Ex. 2146
`IPR2023-00922
`Page 4 of 17
`
`

`

`TAT/2017
`
`Hiroshige Goto's Weekly Foreign News
`
`Backward compatibility with PS is one thing, but I want to
`support everything from dirty legacy (graphics) to the latest
`shader. We want to whip out a resolution of WSXGAorhigher.
`Whenthat time comesit will faster to bring everything (from
`NVIDIA)all at once than for us to build it from scratch.
`
`{Q] Microsoft used a Unified-Shader architecture for ATI's
`GPU in the Xbox 360. Is Unified-Shader more advanced in
`
`programmability?
`
`{Kutaragi] ATI's architecture looks good atfirst glance
`because the shaderis shared andits architecture is identical to
`
`the Vertex Shader and the Pixel Shader, but I don’t think so.
`For example, where do you put the results of vertex processing
`and how do youflow it back to the shader(for pixel processing).
`The whole thingstalls if any part of it gets clogged up. The
`picture that is being colored is different from the reality. When
`it comes downto practical efficiency, I believe NVIDIA's
`approachis superior.
`
`«Compatibility is maintained
`through a combination of
`hardware and software.
`
`[Q] Will compatibility with older
`PlayStations be achieved with
`hardware?
`
`
`
`PLAYSTATION3
`
`{Kutaragi] That will be done through a combination of
`hardware and software. If you want to do it with software alone
`there are many waysto achieve compatibility, but what’s
`important is how close we can get to perfect compatibility.
`
`The people who develop our software do unexpected and
`unimaginable things. For example, there was a program that
`wasn't logical but just so happened to work. There are cases
`where code is working,but it’s working for a completely
`different reason than what was intended. There are times when
`
`a code passes our testing and weare like, "whatis this code!?”
`
`Weneed to have compatibility with coding like this. Butit's
`not logical, so it's a bit of a hassle (to achieve compatibility
`using software alone). Sometimes hardware is needed, We can
`rely on
`
`http://pe watch impress co.jp/docs/2005/0613/kaigail89 htm
`
`4/8
`
`AMD1318_0157005
`
`ATI Ex. 2146
`
`IPR2023-00922
`
`Page 5 of 17
`
`ATI Ex. 2146
`IPR2023-00922
`Page 5 of 17
`
`

`

`TAT/2017
`
`Hiroshige Goto's Weekly Foreign News
`
`hardware in some areas and software in others with the power
`of the PS3.
`
`[Q] Whatis the CPU's endian when emulating the CPU-side
`code with software.
`
`{Kutaragi] Cell is bi-endian, so it doesn’t matter.
`
`[{Q] Xbox 360 will be compatible with mainly just software.
`They don't have any other choice because they don't
`manufacture their own chips, what do you think about this?
`
`{Kutaragi] The current Xbox will be older generation when
`the new generation comes out this November. In doing so,
`Xboxis essentially wiping itself out. The only way they can fix
`this is by having 100% compatibility from day one. But
`(Microsoft) probably won't be able to commit to that because
`it's a technical difficulty.
`
`¢SCEI and NVIDIAare both very similar
`
`Kutaragi's answers suggest that the relationship between
`SCEI and NVIDIA goes beyond a simple deal to develop
`devices; there is also the element of shared corporate direction
`and culture. Both companies like exploring novel ideas, take
`risks, and push costs to pursue functionality and performance.
`Thatis not to say that they always dothis, but it is a tendency.
`Both companiesare also in agreement on the idea of focusing
`on processors now.
`
`NVIDIAis particularly oriented toward programmability
`compared to other GPU vendors. Specifically, ATI
`Technologies and 3Dlabsalso have a strong focus on
`programmability, but NVIDIA has been the most aggressive in
`improving versatility. NVIDIA will increase the GPU's die size
`(the area of the semiconductor) to achieve this, and are
`prepared for manufacturing cost to skyrocket. It appears that
`SCEIthoughtthat the direction NVIDIA is going in made them
`suitable as a partner.
`
`At the moment, GPUsare rushing toward becoming
`programmable processors that specialize in stream processing
`(processing large amounts of data in small program pieces in a
`
`http://pe watch impress co.jp/docs/2005/0613/kaigail89 htm
`
`5/8
`
`AMD1318_0157006
`
`ATI Ex. 2146
`
`IPR2023-00922
`
`Page 6 of 17
`
`ATI Ex. 2146
`IPR2023-00922
`Page 6 of 17
`
`

`

`TAT/2017
`
`Hiroshige Goto's Weekly Foreign News
`
`stream-like manner). The system will be able to perform more
`general processing beyond graphics by improving the
`versatility of programmable shaders, i.e. the arithmetic core.
`On the other hand, the main idea of Cell is for it to be a
`general-purpose processor with sub-processors optimized for
`stream processing. The general-purpose processoris being
`evolved into a structure suitable for stream-type processing,
`something that will become more important in the future. In
`other words, SCET and NVIDIA are working towards the same
`goal from different directions and giventhis it is not surprising
`that SCEI and NVIDIA have shared visions. It is also
`
`understandable that the goals of the two companies match in
`terms of the ideal processor.
`
`Kutaragi's answers confirm to us that several options were
`considered for the PlayStation 3 graphics architecture. First
`was the idea to have a single Cell processor perform graphics
`processing. Cell's data processing processor core, SPE,is
`equipped with a SIMD(single instruction, multiple data)
`arithmetic unit, which can essentially do the same thing as a
`programmable shader, which also has a SIMD structure.
`However, it is not realistic to have a Cell perform graphics
`because it would naturally reduce the performance of the Cell
`asa CPU,
`
`The second option was to install two Cells and use one of
`them exclusively for graphics. It can be assumedthat this idea
`involved a proposal to expand Cell architecture into graphics
`and use the SPE for graphics processing. In this case, I assume
`that they would have installed an arithmetic unit specifically for
`graphics processing. However, plans for a Cell-based graphics
`chip are said to have ended at a very early stage.
`
`Incidentally, current PS3 architecture can still use Cell for
`graphics processing. NVIDIA's David B. Kirk also revealed
`that 3D graphics pre-processing and post-processing can be
`performed in a Cell's SPE with a Cell/RSX combination. For
`example, displacement mapping, which transforms vertex data,
`can be performed by the SPE.
`
`http://pe watch impress co.jp/docs/2005/0613/kaigail 89 htm
`
`6/8
`
`AMD1318_0157007
`
`ATI Ex. 2146
`
`IPR2023-00922
`
`Page 7 of 17
`
`ATI Ex. 2146
`IPR2023-00922
`Page 7 of 17
`
`

`

`TAT/2017
`
`Hiroshige Goto's Weekly Foreign News
`
`The reason why SCEI did not put eDRAM (embedded
`DRAM)on the RSX is obvious from the supported screen
`resolutions, as I mentioned before. It is also clear that the
`company did not want eDRAMto useup die area in order to
`achieve high shader processing performance. This is a
`fundamentally different approach compared to the PS2's
`graphics synthesizer, which took advantage of the wide
`bandwidth of eDRAMto create special graphics architecture.
`Based onthe information available on the RSX,the architecture
`is colored by NVIDIA’s involvement.
`
`SCEIsolved the compatibility issue in the PlayStation 2 with
`hardware by installing the old PS chipset as a sub-processor.
`That was becauseit’s impossible to ensure almost 100%
`compatibility without using hardware-based solutions.
`Emulating hardware completely in software requires an
`enormous amount of CPU power. This is especially critical for
`machineslike the PS2, where the content of the hardware is
`made public so that developers can freely access resources.
`
`Whatis clear now is that they are essentially trying to
`achieve near-perfect compatibility with the PS3. Therefore, this
`direction of hardware-based compatibility will continue with
`the PS3. This time however, they will make use of Cell's high
`processing power andalso have software-based compatibility
`(emulators). The fact that they wentto the trouble of making
`Cell bi-endian meansthat they will have compatibility on the
`CPUside through Cell. In the early stages of co-development
`with IBM, SCEI reportedly requested that bi-endeian be
`required for compatibility. Incidentally, this time compatibility
`will be achieved for two generations; both the PS and PS2.
`Both the PS and PS2 have MIPS architecture CPUs.
`
`
`
`
`
`Related articles
`
`[June 8] [Overseas ] Interview with SCEI President Kutaragi (1)
`Transforming Computing with PlayStation 3 |
`
`http://oc. watch.impress.co.jp/docs/2005/0608/kaigai 186.htm
`[June 9] [Overseas ] Interview with SCEI President Kutaragi (2)
`“PS3 HDD with full-featured Linux”
`
`http://pe.
`watch.impress.co.jp/docs/2005/0609/kaigai187.htm
`
`
`
`Back issues
`
` http://pe watch impress co.jp/docs/2005/0613/kaigail89 htm
`
`V8
`
`AMD1318_0157008
`
`ATI Ex. 2146
`
`IPR2023-00922
`
`Page 8 of 17
`
`ATI Ex. 2146
`IPR2023-00922
`Page 8 of 17
`
`

`

`17/2017
`
`Hiroshige Goto's Weekly Foreign News
`
`(June 13, 2005)
`
`[Reported by Hiroshige Goto]
`
`
`
`[PC Watch homepage]
`
`
`PC Watch Editorial Department: pc-watch-info@.impress.co.jp We do not respond to
`individual questions.
`
`Copyright (c) 2005 Impress Corporation, an Impress Group
`company. All rights reserved.
`
`http://pe watch impress co.jp/docs/2005/0613/kaigail89 htm
`
`8/8
`
`AMD1318_0157009
`
`ATI Ex. 2146
`
`IPR2023-00922
`
`Page 9 of 17
`
`ATI Ex. 2146
`IPR2023-00922
`Page 9 of 17
`
`

`

`THF2017
`
`HYLROWeahyBHoOA—-A
`
`m48FR, 3ODWeeklytz:
`ss=1—Aa
`
`SCElIAS RAKHETIUIE
`1.—(3)
`[PSITCNVIDIALEMAE
`FH J
`
`YIRRALBDORF hi
`
`els.
`
` eCellIN-ADISIA YIRF eaayee
`
`IF
`CQ) Ps377-FF9F VTC.
`SRUIEOISIS Da vIOAM
`CelIN-ADIP?—-FFIFVCE
`DOECEE,. Bt. CellXN—A
`DGPUZHE SRDDEON.
`
`AF BAER
`
`CAFRAK] CellD7(HOSPE(Synergistic Processor
`Element) 2959 4 vOA(CHSTHSACTES. KH.
`ESCOFEOC DIME. D577 yvoa27Ory oH
`FLUVARMTC. CoMEIFCLDSI ULDIECOIS I 414
`PADS T#EUTWIS. LL. CAWUSHUAIS. G
`DIELWVRL. Celllcld, BI CHhITWPSCEMBS.
`
`Colle2AT. (AADCAIRIDI4 YVOAPtAT
`MOcCWSRKGESVEM, D> EAa-—VFEUTOCel
`&. Shaderé LUCOMARE(SMADIES EBAT. ik
`te. Shaderl$ShaderC. MMEKITIIS I-74 v OAC)
`LE, BSDSCEMCESP—-—FFIVFVICULELY
`Hoé6. MAIL. FerRAAPL—-Avy CLA(ASPET
`FRD)EWVETEBCES.
`
`CONEC(OUPILAT A3DGS 74 YIAjs. EH
`SL<ERTE. BRICGDIS I+ YIA)SHOPTC
`(335. EVCEH. SECORBETIED De. SA
`
`httpd/pc. watch im press. co.jp'dars/2S061k gai 189.htm
`
`Ve
`
`AMD1318_0157010
`
`ATI Ex. 2146
`
`IPR2023-00922
`
`Page 10 of 17
`
`ATI Ex. 2146
`IPR2023-00922
`Page 10 of 17
`
`

`

`THF2017
`
`HHYLROWeahyBHoIA—-A
`
`DHFRCH. Xbox 360TH ThISF-—AMECA EM,
`ESLE3DE.
`
`LOL, R5(k, SSA C(BRGENBDBAOPCR
`PRRAMNS3DIC)LIEL\. EORMI(clk, AKER O(CPU
`éGPUCT-PFRKALEWEWSBANHS. IED
`5. SEOrP-FFOF VeERoke. ARESGPUL
`COPE PRAI— wy Kit, HEDSDSO>KA)
`A(CESECINTAUICULZL. SHI. AVRO
`K.EAE—-SHICD TUS. EDS, BHL\ MA
`CGA ReMESCEMCES.
`
`eJJLHDIVORM(ccDRAMEPHOS
`
`(Q]) BYSI4 YvIAZRAE (heDRAM(AAA+AG+DRAM)
`PeTUES. HDTV 2H ¢BlYT. eDRAMICL
`D7 EH ASMECRE.
`
`(ABRBAK] ARK. GPUICISI47 vyIRZAEUIMB
`< TB. Redwood(Cell ERSX#RMS SRBT OYI— 7
`XITA)EYDRAM(XDR DRAM®O I — FR—A)CHIE
`C&S. YDRAMIALAIP-T KECK EV CD TNS
`5.
`
`€35lal\9 Ch. (GSI 4 vyIZAO)HBAAN
`LYL,
`ISLEYShader( VE a —ra SD) TSHBIC. UBL
`BM(ODAEVICPVILCATSCE/)C. (HRMEDT AL
`BF LQ\LACUTUWOD EWS MESS. thon’
`COCaAMDWO(AE UMA. BPKLORZDCESY
`(LAL. Shader{dShaderC. HOT CUBORBATS
`D5, ES55(COAEVOMEHMHS. FiO. DIL
`HDTV, 2kx1k(1,920«1,080 KY He FOIL vy TIe
`im TAT ARL WEL ERBSE. KEOVRAMMYRER
`Be
`
`eDRAMISHE. eDRAMEHSOIS, PS2
`€3H3B&.
`OPAL. CH. SAMMICHPOR. 6
`L.
`(HDTV#DAK— bh CES BO)eDRAM*#200¥Ammm
`£D300FAmc DOAORCANEETS. CSL
`5.
`(eDRAMOMRORHOICF yA(THRCESOTY
`IBS 3 HRD TLL. ShaderMORARo TLEHS.
`
`httrdpc. watch im press.co.|pdars/205/061 Yaigai 189.hon
`
`2B
`
`AMD1318_0157011
`
`ATI Ex. 2146
`
`IPR2023-00922
`
`Page 11 of 17
`
`ATI Ex. 2146
`IPR2023-00922
`Page 11 of 17
`
`

`

`THF2017
`
`HYLROWeahyBHoOA—-A
`
`ENnAkODB. WvcOVyZict#o T. << &AShader
`ERETETIL.
`
`NVIDIACHASTSHHOFORYvOrVvaYy
`
`€EEB.
`([Q]
`ATEON'.
`
`BEGPUADS—-—OPfetNvIDIAcé
`
`[LASRAKR] CNET. Ret. DYEa-sLTy
`FIVRXUbhAONVSI*AVIRAZSR, BSAA Hie
`HS5P 2 TER. FOVARWEAOT. ECeCAPD
`jz. CUT. SO. IDV FA-PENSEOEPSEHIC
`NVIDIA & #8 AE.
`
`NVIDIAISAPCIDS D4 VORZEETEABRET,
`IntelRPORYGDPo TVET ERT. Programable
`ShaderTPS SEUTIIS. NVIDIADAFOR YtEULT
`OBERETTERE IBR I SOA, FEF - H—DDavid
`B. Kirk,NVIDIA@ChiefScientisN R22. FASEBD'SGI
`BEAKSERI>Y LPI-GSREOWSFEDSE. K
`Blk, FyPOKESREMSMCeP(c, PORTE
`BUG OIERUEWNECWSERDES. We. POT
`ZSCLEBHSSM, DILFv (HE SICUTUS.
`
`NVDIADVPIO-F ¢cRSOrYP7O0—-Fld, meHC
`ASWLIOISVINOPFORV VGH, ETKAIARE
`KDEWSAC-RALECWNS. VEY>(Jen-Hsun
`Huang. NVIDIADILRFECEO)NPT EY REL<K METS
`BEMSSM, CORICHSOM, BROTFOtYt?
`ADKEWUSG. BMEWISOSL. 4H. REMPC. 1
`~, REDE TOPOVYVseRBrASTORYvOI
`fe
`
`KSSCACAEOAMITADDTHO, EDBKT
`id, RIEEDALFHKALTIIS. O-FVYy TBH
`ALTWS. KK. HSlk RIOP-FFIFVPDS
`SMBZRITUIS,. SAMOAMNTT, WOT
`CAPOREWEBATISDS, NVIDIACHAL.
`
`HIVE DOBRA. FLRAZTFLIMBEEIVLA
`GRRREICBODDESA. BREECORILAICIOL
`5. TVBPCH. SCNRTMBST SiRICRS. TENS,
`SANTHHRICOAR— KUL.
`
`httpdipc. watch im presa.co.|pidars/215061Ykalgal 189.hon
`
`338
`
`AMD1318_0157012
`
`ATI Ex. 2146
`
`IPR2023-00922
`
`Page 12 of 17
`
`ATI Ex. 2146
`IPR2023-00922
`Page 12 of 17
`
`

`

`THF2017
`LMLROWoalyRH—A—A
`PSEOD FURERESESEL, HULALD—-(DIS79
`4 YIA)YSRHOShader*XCHMAK— KOE.
`MEISWSXGAU LODEOH/Sy eEthbel.. ESL
`Tet. RAMBRIDVF(EIL EY CHSAO, (NVIDIAN
`S)AMANS co TARAHU).
`
`([Q] Microsoftit. Xbox 360T(ILATIOGPUMUnified-
`Shader7? —-+>O7F vehok, POGSAtt tlt
`Unified-ShaderOAM 5cHy Tid.
`
`CAPRAK] ATIOP—-+TOF vid, ShaderldseH
`CVertex Shader &Pixel Shader?" A—(7 —FF7-OFv)C.
`—-RACKRASIEC. BUNCE BS. WAL. KALB
`MUIREETICEC OD, CHESS TShader
`(C(ETRILABOEMDC)BURHTON. ECDDBE
`St, SHAKILL TUES. BIcHlWEbEDOERE
`CSRS. MANGMRBEZBSANIL, NVIDIADV TO
`—FOAVENTISENS.
`
`oHREDEGIZN—KOLPE |
`YVIFILIPORSEDST
`
`CQ) MEOPlayStationt¢DLik
`fHEID/\-— FOL PYCHRITSO
`
`Dd’.
`
`PLAYSTATION 3
`
`CAPRARKI J\-KOLVEVYD ROLYOIVER
`—P3BYTRS. (VIFILPREITOPSSCBAlE
`EDICTHBSM, ENIEITTER ICUVEIRPEICIBL VA
`DBOVKHF.
`
`YIKEMRULTUSAILRIME. MRCERICE
`&RPDTLES. MA. POTSLEL CMH TH
`USE, EREKMUECWDE. MUTUSUTE, TF
`BEML CHOMACHUTEEWNSKSRT-AN
`DD. RIEDFAKBK CORT.
`[ATETOI-K
`(411 HELI— EMM TCLESBAMSS.
`
`R(t, SSULED—RieomMTSRREBRSB<KT
`(SUT. LAL. RBUPRUDS,.(VIbULP
`EIFSCHRERMSOIHNS&k TD ¢HLLW J\-— KOUe
`{CHD TC SOEHS. TH. SEPS3)K SUD/I—
`
`httpd/pc. watch im press.co| p'dars/2S061kaigal 189htm
`
`48
`
`AMD1318_0157013
`
`ATI Ex. 2146
`
`IPR2023-00922
`
`Page 13 of 17
`
`ATI Ex. 2146
`IPR2023-00922
`Page 13 of 17
`
`

`

`THF2017
`
`HYLROWeahyBHoOA—-A
`
`Mans SBSETCAIA\—FCHSETCSALVIKE
`CLUDE HUCES.
`
`[Q] CPUMIDI-—F#YID FCLE12L— kTSHF
`{(ClACPUMDLYF474 7 Ha.
`
`(CABS BRAK) Celt)HLF41PVEMS, 3c
`CHS.
`
`CQ] Xbox 360Id(FIFU OD MOL PESCEREER
`So. RMSld4,. FYPHAHCMM TLVALIOS, falc
`IRDRUDI, EDHSON.
`
`CAS RAK] Xboxid, HiROSEONAICRS
`&, BUFOXboxiSIAttRi(cAS. CSTSE. XboxltH
`SCBPERLETLES CEILS. ENEKSHE—-D
`FSOVDAKRAZVYADSMSCKE. CE
`(MicrosoRhISNENAI Sy khCHBAUESS. HAAHIC
`6H UL.
`
`eSCEIENVIDIAISMESOR+
`
`RB RAKROBGRD SIS, SCEIENVIDIADMUDS
`lcld, BMICFIHTARRORSULO. EREOA mitt
`PPILF P—D-HEWSBRVSST EMMENIT<
`SB. Med. MRRPAFPVeA. UROFAN-
`C. DTARFOZOR CHAMERTIERTS. Wek
`WSDITTit, ESLER. KE. Mth
`¢6. HESPTOVYSeBRISREC-RET
`De
`
`GPUNLYS—OFTC. NVIDIASHICPODSVIILE
`AOE ETEDSRU. IEH(C BAIE. ATI TechnologiesY?
`3Dlabs6 7 OFSVAIUCIC PALBAH TUISOIE
`DS, NVIDIADUAE HOS CHC-BPILY TIER
`Diz. NVIDIAIS. EDEMIT. GRUODD1H1A
`KARODUEKLL. RHIAKYRRTS CEB
`Hash). SCEIEUTISZ. TSUENVIDIADIS AE
`DS. M— BFK ELTAWTUSEBRARKSE.
`
`RE. GPULAKU-APORvy DT TIAhee7o07
`ShLeE-ACKBT -—SHAZ KU DRITMBL CH <)
`(CHE LIEP OPS VINTF OR YONA ERHEL TU)
`
`httppc. watch im press.co.|pdars/205/061 Yaigal 189.hon
`
`8
`
`AMD1318_0157014
`
`ATI Ex. 2146
`
`IPR2023-00922
`
`Page 14 of 17
`
`ATI Ex. 2146
`IPR2023-00922
`Page 14 of 17
`
`

`

`TATROI7
`BAVADWehyRea2.
`Bo RAASY’CHSProgramable ShaderOUMEZ BOS
`TET. PSILVIAUHO. LOWKHRUMBT
`SSLSICTSNMHE. —H. ClORABMB. ZAR
`U-L7OVv> TCRDEGIF ORY VER
`BERADPOCH ENS PHIPreE. ART ORY
`RHEMLEEC. SRBBCRSAZAbU-ARTOtY*S
`VIC RUEMIC TS. Bo CHMUE, SCE
`NVIDIA(S, (ACT-ULICaIaDS PI O-FUDD
`HS. COBASBEK. SCEHENVIDIAM ET AYCR
`LEEWSOS ABBRC ETL), GAOA RHE
`(c. BHOTOtyéE\5S-KADSSTKLEBHT
`#S.
`
`AB RAKOMMD Sik, PLAYSTATION 30757 -4
`VYIRP - EFOF? TAS DORR RH LIE
`LEMBMTISENS. HT. OcOryye, 7
`S74 YIRZMBEP SHS. ClOF—-FUBIO
`Py?CH SSPEls. SIMD(Single Instruction,
`
`MultipleDataOweI—y heHATHEO. AO<
`SIMD tOProgramable Shader CHER MIT(AROT EN
`CaS. LOLA CelIMICISI4 VOREPSRSTE
`
`is. 4ROTERDNS, CPUEL THCEIMOTEMES Alo
`CTCLESED, BRATCARL.
`
`HC. CelleAMBRE. HTOCHWEISI4 97
`ABMICRSR. CORICld. PSD1 VORAMNCedMD
`F-FFUFVERT. SPEXDS D4 vyORZMB
`AISICTSRESENCLVELCHEENS. EDF
`(4. PSI47 vOZOHZUBAIORBITY how
`RISEBTSDNIIESD 3 CRUENS. 52 <8.
`CIN 2OIS D4 VORAFVIOTS MA DRO
`DYDRRCHAEEADNTUG,
`
`BSAC. RIEOPSIP—-FFOF VCH. ClheIS
`74 YOAZRAHES TE(ATAS. NVIDIADKikK
`t£, CeERSXDBAASDECA IDISI4YIZD
`FUTFOV YT VIOPRA hT ORY YDPRCIDSPE
`CHBASCEBMSINCLTWIS, MAI BAS
`SREINTSDisplacement Mapping(F 4#AFELTARXY
`KVVELDAERSPEN TITAS CEBCHS,
`
`httpdips watch im press co[placeAIG041 Yaaga'188 horn
`
`SS
`
`AMD1318_0157015
`
`ATI Ex. 2146
`
`IPR2023-00922
`
`Page 15 of 17
`
`ATI Ex. 2146
`IPR2023-00922
`Page 15 of 17
`
`

`

`THF2017
`
`LHLAOWeehyRHIA—A
`
`SCEIARSX lc eDRAM(S8GHIAG+DRAM) CRB ILDFE
`Fit. LRICBSBUEESBOTOM-bKTSHHRR
`BYSWMEK. He. Ml iShaderFSOVyL VINA
`—VVAERRISEOI(CIs. eDRAMCA1HRA
`FSDIHCEARWEBARTEBIYS. CM.
`eDRAMOJZSRI LURID D4 VORP
`FOF? ZR IE. PS2OGraphics Synthesizer¢& (SHA HY
`(CHMBM AGS. RSXCAMENTUSHRERSR
`DN, P-FFOF vMlclt. NVIDIAB DHHSTL.
`
`SCEI(4. PlayStation 2T(SIHPSOF YPVy KET
`FORCVYGPELTHRISCETAN-KULPICHKIT
`HRMOMMARRRELR. Child, J KROLPA-R
`(CLUE,
`[FFOO%VOFAEAHARTERUIDSIE.
`A-BFULYVeESICVIBOULYCLEILVary
`SSE. BKACPUNI-DUELERS. Child. PS2D
`ADICA-— FOLIOPSEAAUT. WREM AIC
`YY—-AlPIVACTESEBICLEVE SCS. HIT
`DUF AVIV.
`
`REWSMWUO(S. PSICHRAM (CSB LVAR
`HERRRLADSELTIISCER. EMDEH. PITH
`RE UTA BOL PR-ACO Hiatt CLD Halt
`MRANS. EL. SElk. CMORAOVRYTLTID
`NO-RIDUT. YVIRDLPR-ATORIMA(LE1
`L—S)bRS5NS. NDENSCAR SF LYF414 Pvc
`LEEWS TEICellTCPUTYH FOGRHARSC
`C#HERKTS. BMCORMMREOMMRAC. SCEMS
`HAEORMIOILF 4 PTDWREeEBRLEC
`{AS5NS. HSthdsl<=. SAIS. PS. PS2E2HNHOHM
`tETRRITS. PSkPSAMMA CBMIPSV—FTIF 7
`DCPUCH RLTID.
`
`ob idee St
`(6A8H) (#5) SCEHIAS RARRT>OEI—(1)
`[ PLAYSTATION 3C JV EI —F 4 VDOREAS)
`http://pc.watch.impress.co.jp/docs/2005/0608/kaigai 1 86.htn
`C6A9A] Cob] SCELASRAKHERTY>SEA—(2)
`[PS 3QHDDIc FD) LBBHELinux25a |
`http ://pc.watch.impress.co.jp/docs/2005/0609/kaigail87.htm
`
`ol \ ywOts)\ ae
`
`
`
`httpdipc. watch im press.co| p'dacs/ZIS0613agal 189.hon 78
`
`AMD1318_0157016
`
`ATI Ex. 2146
`
`IPR2023-00922
`
`Page 16 of 17
`
`ATI Ex. 2146
`IPR2023-00922
`Page 16 of 17
`
`

`

`THF2017
`(200544613 A)
`
`HYLROWeahyBHoOA—-A
`
`[Reported by2##5/7¢@linoshigeGoto))
`
`PC Watch7t—_AA—F) __
`
`PC WatchBRAYpe-watch-info@impress.co.jp 7PATIL
`TC. MUNCTBIFAAUE LE BA,
`
`Copyright (c) 2005 Impress Corporation, an Impress Group
`company.All rights reserved.
`
`httpd/pc. watch im press.co.| ptdars/21050613kaigai 189ht
`
`AMD1318_0157017
`
`ATI Ex. 2146
`
`IPR2023-00922
`
`Page 17 of 17
`
`ATI Ex. 2146
`IPR2023-00922
`Page 17 of 17
`
`

This document is available on Docket Alarm but you must sign up to view it.


Or .

Accessing this document will incur an additional charge of $.

After purchase, you can access this document again without charge.

Accept $ Charge
throbber

Still Working On It

This document is taking longer than usual to download. This can happen if we need to contact the court directly to obtain the document and their servers are running slowly.

Give it another minute or two to complete, and then try the refresh button.

throbber

A few More Minutes ... Still Working

It can take up to 5 minutes for us to download a document if the court servers are running slowly.

Thank you for your continued patience.

This document could not be displayed.

We could not find this document within its docket. Please go back to the docket page and check the link. If that does not work, go back to the docket and refresh it to pull the newest information.

Your account does not support viewing this document.

You need a Paid Account to view this document. Click here to change your account type.

Your account does not support viewing this document.

Set your membership status to view this document.

With a Docket Alarm membership, you'll get a whole lot more, including:

  • Up-to-date information for this case.
  • Email alerts whenever there is an update.
  • Full text search for other cases.
  • Get email alerts whenever a new case matches your search.

Become a Member

One Moment Please

The filing “” is large (MB) and is being downloaded.

Please refresh this page in a few minutes to see if the filing has been downloaded. The filing will also be emailed to you when the download completes.

Your document is on its way!

If you do not receive the document in five minutes, contact support at support@docketalarm.com.

Sealed Document

We are unable to display this document, it may be under a court ordered seal.

If you have proper credentials to access the file, you may proceed directly to the court's system using your government issued username and password.


Access Government Site

We are redirecting you
to a mobile optimized page.





Document Unreadable or Corrupt

Refresh this Document
Go to the Docket

We are unable to display this document.

Refresh this Document
Go to the Docket