Uma ungajwayelene nokusha I-Quadro RTX line, nakhu okudingeka ukwazi. Amakhadi amasha e-RTX asebenzisa isizukulwane esilandelayo sama-GPU asekelwe ku- Ukwakhiwa kwezakhiwo izici ezihlangana ndawonye ku Ipulatifomu yeNVIDIA RTX.
I-Turing yengeza izici eziningi ezibalulekile ngaphezu kwezakhiwo zangaphambilini, okuyinhloko okungama-Tensor cores amasha anezela amakhono okufunda ajulile e-AI kanye nama-RT cores okusheshisa ukulandelela umsebe wesikhathi sangempela. Inkundla ye-RTX ihlinzeka ngama-RTX API ahlukahlukene konjiniyela okubavumela ukuthi bakhe izinhlelo zokusebenza ze-RTX ezisizakala ngezakhiwo ze-RTX Turing futhi ikuvumela ukuthi usebenzise lezo zici ze-RTX ezinhlelweni ezisizakala ngomugqa we-NVIDIA wama-RTX GPU. Ukutholile lokho? Kuhle.
Ukukufingqa, ama-NVIDIA RTX GPU angakwazi ukwenza okungekho amanye ama-GPU ngaphambi kokuthi akwenze. Kuze kube manje, kubangele uguquko embonini yemidlalo kanye nesoftware ye-3D, enjabulweni yabathuthukisi abathile kanye nokudumala kwabanye okwamanje, okwenza kwaba nokwenzeka nge-GPU ukuphatha ubunzima bokulandela umsebe wesikhathi sangempela obekwenzeka kuphela. ku-CPU ngaphambili.
Ngenxa yalokhu, amakhadi e-RTX, anama-RT cores kanye nama-Tensor cores, ethule izinga elisha lokusheshisa isoftware + hardware. Ngokuhlangene imiphumela yokulandelela imisebe yesikhathi sangempela yayifana nento engakaze ibonwe ngaphambili, kukho kokubili isivinini nokusebenza.
I-NVIDIA njengamanje inezinketho ezinhlanu ze-Quadro RTX - i-RTX 3000, i-RTX 4000, i-RTX 5000, i-RTX 6000, i-RTX 8000, ne-RTX 4000 ehlezi njengendawo yokungena kulo lonke iqembu.
I-Quadro RTX 4000 Imininingwane
Manje, i-Quadro RTX 4000 akuyona i-GPU enamandla kakhulu endaweni ye-RTX kodwa ifika endaweni emnandi enephuzu lentengo khona kanye ngaphansi kwe-$900. Sike sabheka i I-Quadro RTX 4000 esihlokweni esandulele esichaza imininingwane yayo. Ukuze unikeze ukuvuselela okusheshayo nakhu ukuqhathanisa izici phakathi kochungechunge lwe-Quadro RTX.
| CUDA Cores | Ukuqhathaniswa kweQuadro RTX GPU | |||
| I-Quadro RTX 4000 | I-Quadro RTX 5000 | I-Quadro RTX 6000 | I-Quadro RTX 8000 | |
| I-GPU Memory | I-8 GB GDDR6 | I-16 GB GDDR6 | I-24 GB GDDR6 | I-48 GB GDDR6 |
| I-Bandwidth yememori | I-416 Gbps | I-416 Gbps | I-624 Gbps | I-624 Gbps |
| CUDA Cores | 2304 | 3072 | 4608 | 4608 |
| Ama-Tensor Cores | 288 | 384 | 576 | 576 |
| Ama-RT Cores | 36 | 48 | 72 | 72 |
| I-RTX-OPS | I-43T | I-62T | I-84T | I-84T |
| Imisebe yabalingisi | I-6 Giga Rays / Sec | I-8 Giga Rays / Sec | I-10 Giga Rays / Sec | I-10 Giga Rays / Sec |
| Ukusebenza kwe-FP32 | I-7.1 TFLOPS | I-11.2 TFLOPS | I-16.3 TFLOPS | I-16.3 TFLOPS |
| Uhlu lwamanani | $900 | $2300 | $6300 | $10000 |
Futhi uma usekhadini elidala le-P(ascal), i-K(epler), noma i-M(axwell) yochungechunge lwe-Quadro, ngeke ube nama-Turing RT cores kanye nama-Tensor cores ahlinzeka ngomkhondo we-real-time ray namandla we-AI. kodwa uzothola ikhadi elinenani eliphansi. Nasi isiqhathaniso se-RTX 4000 namakhadi ochungechunge lwesizukulwane esedlule:
| CUDA Cores | Ukuqhathaniswa kwe-Quadro Mid-Range GPU | |||
| I-Quadro RTX 4000 | I-Quadro P4000 | I-Quadro P2000 | I-Quadro P1000 | |
| I-GPU Memory | I-8 GB GDDR6 | I-8 GB GDDR5 | I-5 GB GDDR5 | I-4 GB GDDR5 |
| I-Bandwidth yememori | I-416 Gbps | I-243 Gbps | I-140 Gbps | I-80 Gbps |
| CUDA Cores | 2304 | 1792 | 1024 | 640 |
| Ama-Tensor Cores | 288 | - | - | - |
| Ama-RT Cores | 36 | - | - | - |
| I-RTX-OPS | I-43T | - | - | - |
| Imisebe yabalingisi | I-6 Giga Rays / Sec | - | - | - |
| Ukusebenza kwe-FP32 | I-7.1 TFLOPS | I-5.3 TFLOPS | I-3.0 TFLOPS | I-1.9 TFLOPS |
| Uhlu lwamanani | $900 | $750 | $399 | $319 |
I-Quadro RTX 4000 Ukuqhathanisa Ukusebenza
Lokhu ngeke kube ukubuyekezwa okuphelele kokusebenza kwe-RTX kodwa bengifuna ukunikeza ukuqhathanisa phezulu komugqa we-P-Series GPU abaningi okungenzeka basenawo futhi, empeleni, kusebenza njengomugqa phakathi kwe-NVIDIA GPU yesizukulwane sangaphambilini ne-NVIDIA. Ama-RTX GPUs.
Uhlelo lwethu lokuhlola luyi- Indawo yokusebenza yeLenovo ThinkStation P520. Ihlala phakathi nomndeni wesiteshi sokusebenza seLenovo P-Series. Ukucushwa kwesisekelo ngenketho ye-Quadro RTX 4000 kungena ku-$1900. Bona wethu Ukubuyekezwa kwe-ThinkStation P520 ukuthola eminye imininingwane futhi ungayishintsha kanjani i-GPU uma udinga ukuthuthukisa ikhadi elidala.
I-P520 ifike neQuadro P4000 enekhono kakhulu. Ikhadi levidiyo elihle kakhulu elimaphakathi nebanga, ikakhulukazi ngentengo. Noma kunjalo, uma kunikezwe i-P4000 yaphuma ngoFebhuwari 2017, i-RTX iqinisekile ukuthi izothuthukisa ukusebenza, noma zombili zisebenzisa umshayeli we-Quadro GPU wakamuva.
NjengeLenovo P520, kunezindawo zokusebenza eziningi ze-3D CAD ezine-P4000 njengenketho yokwenza ngokwezifiso okuyisisekelo. Njengoba kushiwo ngenhla, nokho, i-Quadro P4000 ayinawo ama-Turing RT cores kanye nama-Tensor cores amakhadi we-RTX okuthi, ngikholwe, enza umehluko.
Ukuhlolwa kokusebenza kwe-PassMark:
Kunamabhentshimakhi amahle wokusebenza atholakala ngawo PassMark Performance Test esizithandayo ngekhono lokukwazi qhathanisa nama-GPU phakathi kwezinkulungwane zokunye ukucushwa kwesistimu. Ibhentshimakhi yamahhala enikezela ngemiphumela yemisebenzi yekhomputha ehlukahlukene kuyilapho ihlinzeka ngama-avareji namaphesenti. Nansi imiphumela yokuhlolwa kwePassMark.
I-Quadro P4000
I-GPU Compute (Ops./Sec.) : 5526.7 (75%)
I-3D Graphics Mark (Isilinganiso esiyinhlanganisela): 14163.2 (77%)
Isilinganiso se-PassMark (Isilinganiso esiyinhlanganisela) : 6044.9 (77%)
I-Quadro RTX 4000
I-GPU Compute (Ops./Sec.) : 6456.9 (77%)
I-3D Graphics Mark (Isilinganiso esiyinhlanganisela): 17704.6 (90%)
Isilinganiso se-PassMark (Isilinganiso esiyinhlanganisela) : 6161.3 (82%)
Le miphumela iveza i-Quadro RTX 4000 ekwazi ukusebenza okwengeziwe okungu-930 ngomzuzwana uma iqhathaniswa ne-P4000 kanye ne-3D Graphics Mark eyibeka ku-90% percentile, amaphuzu angu-13 ngaphezu kwe-P4000.
Ibhentshimakhi ye-KeyShot:
I-KeyShot Benchmark iyibhentshimakhi yamahhala etholakala kuyo I-KeyShot Viewer. Lena ibhentshimakhi ehluke kancane esebenzisa isistimu yesisekelo ukuze inikeze isilinganiso esisodwa sokusebenza okukhuphukile (noma okwehlisiwe). I inhlukano ichaza amaphuzu angu-1.0 (kungaba yi-CPU noma i-GPU) njengokufana nesistimu yesisekelo, i-Intel Core i7-6900K CPU @3.20GHz, 2601 Mhz, 8 Core(ama). Nansi imiphumela ye-KeyShot Benchmark.
I-Quadro P4000
Umphumela we-CPU: (Intel(R) Xeon(R) W-2133 CPU @ 3.60GHz – Imicu: 12) 1.38
Umphumela we-GPU: (I-Quadro P4000 enomshayeli: 442.50) 3.38
I-Quadro RTX 4000
Umphumela we-CPU: (Intel(R) Xeon(R) W-2133 CPU @ 3.60GHz – Imicu: 12) 1.50
Umphumela we-GPU: (I-Quadro RTX 4000 enomshayeli: 442.50) 23.67
Le miphumela ibonisa ukuthi i-Quadro RTX 4000 ishesha ngo-7x kune-P4000 futhi icishe ibe ngu-16x ngokushesha kunokusebenzisa i-CPU ekusebenzeni okufanayo.
Ukuze uthole ukubuyekezwa okujulile nokuqhathaniswa okubanzi kwamakhadi, ngincoma kakhulu izinsiza ezilandelayo:
- Imithwalo - Ukubuyekezwa kwe-Quadro RTX 4000
- I-Hot Hardware - Ukubuyekezwa kwe-Quadro RTX 4000
- I-PassMark – Isivivinyo sokusebenza se-Quadro RTX 4000
- I-TechPowerUp – Quadro RTX 4000 Imininingwane
Isiphetho
Uma uzama ukunquma ukuthi uzothuthukisa yini ikhadi le-Quadro elidala noma phakathi kwe-P-series Quadro ne-RTX, umehluko wentengo ka-$150 uwufanele ngokombono wami, ikakhulukazi udinga amandla e-GPU ukuze uthole umsebenzi onzima. Lapho wenza ngendlela oyifisayo indawo yokusebenzela ye-OEM, izindleko zokuthuthukisa i-GPU cishe zihlala zinenani elingcono kunalapho ukhetha inketho engenamandla futhi wathuthukiswa kamuva. Ngakho-ke, uma ufuna isistimu entsha, hlola umehluko wentengo ngenketho ye-RTX. Ngicabanga ukuthi uzomangala futhi, lapho uzwa umehluko, ujabule kakhulu ngemiphumela.
Ukudalula: NVI-IDIA inikeze i-RTX 4000 esetshenziswe kulokhu kubuyekezwa futhi iLenovo inikeze indawo yokusebenza. Akukho okokufaka noma ukuhlelwa kwalokhu kubuyekezwa okwenziwe yibo bobabili. UJosh Mings usebenzele futhi njengamanje uhlinzeka ngezinsizakalo ze-Luxion, umdali we-KeyShot Viewer.
