I-NVIDIA isemthethweni evuliwe ikhadi layo elisha lemifanekiso le-GeForce RTX 4070 elihlose ukunikeza i-100+ FPS ku-1440P Gaming ku-$599 US.
I-NVIDIA Ikhulula Ikhadi Layo Lezithombe ze-GeForce RTX 4070, I-Ada Lineup Manje Iqala Ku-$599 US
I-NVIDIA's GeForce RTX 4070 iyisengezo sakamuva ohlwini lwemidlalo ye-Ada Lovelace futhi ithimba eliluhlaza likusho njengokuthuthuka okumangalisayo kubasebenzisi abakhona be-GeForce GTX 1080 kanye ne-GeForce RTX 2070. Ikhadi lezithombe liphinde lehlise intengo yohlelo lwe-Ada Gaming lwehle luye ku-$599 US, okuyikhadi elinenani eliphansi kakhulu lesizukulwane esilandelayo okwamanje. Ngaphandle nje kokungeza ukusebenza okwengeziwe nokusebenza kahle, yengeza uhla lwezici ezintsha ezihlanganisa:
- Ama-Multiprocessors amasha Wokusakaza (SM) - I-SM entsha iletha ukusebenza okufika koku-2x nokusebenza kahle kwamandla
- Isizukulwane sesi-4 seTensor Cores kanye nokugeleza kokubona - Nika amandla futhi usheshise ubuchwepheshe be-AI obushintshayo, okuhlanganisa nezinga lozimele elisha eliphindaphinda i-NVIDIA DLSS 3
- Isizukulwane sesithathu se-RT Cores - Kuze kufike ku-2x ray wokulandela ngomkhondo, iletha imihlaba ebonakalayo enemininingwane emangalisayo kunanini ngaphambili
- Ukuhlelwa kabusha Kokubulawa Kwe-Shader (SER) - I-SER ithuthukisa ukusebenza kokulandelela ama-ray ngo-2x, ikhuphula i-FPS kuze kufike ku-44% ku-Cyberpunk nge-RT: Imodi Yokushayela Okungaphezulu
- I-DLSS 3 - Ukuphumelela okuguquguqukayo emidwebeni enamandla ye-AI ethuthukisa kakhulu ukusebenza kusetshenziswa i-AI ukuze kukhiqizwe ozimele bekhwalithi ephezulu.
- I-NVIDIA Studio - Ukusebenza okungenakuqhathaniswa ekunikezelweni kwe-3D, ukuhlela ividiyo, nokusakaza bukhoma
- Izifaki khodi ze-AV1 - I-NVIDIA Encoder (NVENC) yesizukulwane sesi-8 ene-AV1 isebenza kahle ngo-40% kune-H.264, inika amandla amathuba amasha kubasakazi, abasakazi, nabafonayo bevidiyo.
I-NVIDIA GeForce RTX 4070 isebenzisa ukucushwa okuncishisiwe kwe-AD104 GPU nama-SMs angu-46 ngenani eliphelele lama-CUDA cores angu-5888 I-GPU izofika igcwele 36 MB ye-L2 cache, 184 Texture Mapping Units kanye nesamba sama-ROP angu-64. Ijubane lewashi lekhadi lemifanekiso likalwe kusisekelo esingu-1920 MHz kanye namawashi okuthuthukisa angu-2475 MHz. Ikhadi linikeza ama-TFLOP angu-29 e-FP32, ama-TFLOP angu-67.4 e-RT, nama-TFLOP angu-466 ekukhishweni kwekhompyutha kwe-INT8.
I-NVIDIA ithi ukukhulisa inqolobane ye-L2 izikhathi ezingu-9 ku-RTX 4070 uma kuqhathaniswa ne-RTX 3070 (36 MB vs 4 MB) kuthuthukisa ukusebenza, kunciphisa ukubambezeleka, futhi kukhulisa ukusebenza kahle kwamandla, njengoba ukufinyelela kwedatha kungahlala ku-chip (kunokuba uthembele kumkhawulokudonsa wememori).
Ngokuqondene nama-specs enkumbulo, i-GeForce RTX 4070 inezici eziyi-12 GB GDDR6X ezizolungiswa ngesivinini esingu-21.0 Gbps ku-interface yebhasi engu-192-bit. Lokhu kuzohlinzeka kufika ku-504 GB/s womkhawulokudonsa. Lokhu ukucushwa kwememori okufanayo njengoba kusetshenziswa i-RTX 4070 Ti kanye nememori engu-4 GB ngaphezulu uma kuqhathaniswa namakhadi ehluzo ochungechunge lwe-3070.
- I-NVIDIA GeForce RTX 4070 12 GB TBP “Esemthethweni” - 200W
- I-NVIDIA GeForce RTX 3070 8 GB TBP “Esemthethweni” - 220W
Mayelana nokusetshenziswa kwamandla, i-TBP ilinganiselwe ku-200W. Ikhadi lizonikwa amandla yisixhumi esisodwa samaphini angu-16 esingaletha amandla afinyelela kwangu-600W. Amamodeli angokwezifiso azobe enikeza okuhlosiwe okuphezulu kwe-TBP kodwa futhi avulekele ukusebenzisa izixhumi ezijwayelekile zamaphini angu-8.
2 of 9
I-NVIDIA GeForce RTX 4070 Cards Cards Performance
Mayelana nokusebenza kwalawa ma-GPU amakhulu, i-NVIDIA yabelane ngezibalo zokusebenza kwekhompyutha nezokudlala futhi kubukeka sengathi i-GeForce RTX 4070 izobe ihlezi phakathi kwamamodeli we-3080 10GB kanye ne-12GB ngaphandle kwe-DLSS 3.
2 of 9
Ngenxa yokuqhathanisa nje:
- I-NVIDIA GeForce RTX 4090: 83 TFLOPs (FP32) (2.5 GHz Boost clock)
- I-NVIDIA GeForce RTX 4080: 49 TFLOPs (FP32) (2.5 GHz Boost clock)
- I-NVIDIA GeForce RTX 3090 Ti: 49 TFLOPs (FP32) (1.86 GHz Boost clock)
- I-NVIDIA GeForce RTX 4070 Ti: 40 TFLOPs (FP32) (2.6 GHz Boost clock)
- I-NVIDIA GeForce RTX 3090: 36 TFLOPs (FP32) (1.69 GHz Boost clock)
- I-NVIDIA GeForce RTX 4070: 29 TFLOPs (FP32) (2.48 GHz Boost clock)
- I-NVIDIA GeForce RTX 3080: 29 TFLOPs (FP32) (1.71 GHz Boost clock)
- I-NVIDIA GeForce RTX 3070 Ti: 21 TFLOPs (FP32) (1.77 GHz Boost clock)
- I-NVIDIA GeForce RTX 3070: 20 TFLOPs (FP32) (1.75 GHz Boost clock)
Ngokusekelwe kujubane lewashi elithuthukisiwe elingu-2.48 GHz, uthola kufika ku-29 TFLOPs wokusebenza kwekhompyutha futhi ungakwazi ukuminya okuningi kakhulu nge-overclock njenge sasibonise nge-RTX 4090. Umuntu kufanele akhumbule ukuthi ukusebenza ngekhompyutha akusho ngempela ukusebenza kwegeyimu sekukonke.
I-FP32 Compute Horsepower Comparisons (Okuphakeme Kungcono)
Amandla Okunciphisa
020406080100120
020406080100120
RTX 4090
82
7900 XTX
61
7900XT
51
RTX 4080
49
I-RTX 3090 Ti
40
I-RTX 4070 Ti
40
RTX 3090
36
RTX 4070
29
RTX 3080
29
I-RX6900XTX
25
I-RTX 3070 Ti
22
I-RTX 4060 Ti (Theoretical)
22
RTX 3070
20
I-RTX 3060 Ti
16
RTX 3060
13
I-Xbox Series X
12.1
PlayStation 5
10.2
2 of 9
Ngokuqondene nokusebenza kwegeyimu, i-NVIDIA igcizelela kakhulu ukusetshenziswa kwe-RTX & DLSS3 kanye nemodi entsha ye-RT-overdrive lapho iqhathanisa ikhadi lemifanekiso le-GeForce RTX 4070 nama-RTX 30 GPU amadala.
Ukusebenza kwe-NVIDIA GeForce RTX 4070 1440P:
Ukusebenza kwe-NVIDIA GeForce RTX 4070 DLSS 3:
I-NVIDIA GeForce RTX 4070 eSports Performance:
I-NVIDIA GeForce RTX 4070 Synthetic Benchmarks:
I-NVIDIA GeForce RTX 4070 Power Profiles:
Intengo Nokutholakala Kwamakhadi Ezithombe ze-NVIDIA GeForce RTX 4070
Ikhadi lemifanekiso le-NVIDIA GeForce RTX 4070 lizotholakala kusukela kusasa ngentengo engu-$599 US. Kokubili i-Founders Edition namamodeli angokwezifiso azotholakala ngokuthengiswa ngalolo suku.
I-NVIDIA GeForce RTX 40 Series Imininingwane Esemthethweni:
| Igama le-Graphics Card | I-NVIDIA GeForce RTX 4090 | I-NVIDIA GeForce RTX 4080 | I-NVIDIA GeForce RTX 4070 Ti | I-NVIDIA GeForce RTX 4070 |
|---|---|---|---|---|
| Igama le-GPU | Ada Lovelace AD102-300 | Ada Lovelace AD103-300 | Ada Lovelace AD104-400 | Ada Lovelace AD104-250 |
| Inqubo yeNode | I-TSMC 4N | I-TSMC 4N | I-TSMC 4N | I-TSMC 4N |
| Usayizi we-Die | 608mm2 | 378.6mm2 | 294.5mm2 | 294.5mm2 |
| Transistors | I-76 Billion | I-45.9 Billion | I-35.8 Billion | I-35.8 Billion |
| CUDA Cores | 16384 | 9728 | 7680 | 5888 |
| Ama-TMU / ama-ROP | 512 / 176 | 320 / 112 | 240 / 80 | 184 / 64 |
| Ama-Tensor / RT Cores | 512 / 128 | 304 / 76 | 240 / 60 | 184 / 46 |
| I-Base Clock | 2230 MHz | 2210 MHz | 2310 MHz | 1920 MHz |
| Thuthukisa i-Clock | 2520 MHz | 2510 MHz | 2610 MHz | 2475 MHz |
| I-FP32 Compute | I-83 TFLOPs | I-49 TFLOPs | I-40 TFLOPs | I-29 TFLOPs |
| Ama-RT TFLOPs | I-191 TFLOPs | I-113 TFLOPs | I-82 TFLOPs | I-67 TFLOPs |
| Ama-Tensor-TOPs | I-1321 TOP | I-780 TOP | I-641 TOP | I-466 TOP |
| Imemori Amandla | I-24 GB GDDR6X | I-16 GB GDDR6X | I-12 GB GDDR6X | I-12 GB GDDR6X |
| I-Memory Memory | I-384-bit | I-256-bit | I-192-bit | I-192-bit |
| Isivinini seMemori | I-21.0 Gbps | I-23.0 Gbps | I-21.0 Gbps | I-21.0 Gbps |
| Komkhawulokudonsa | I-1008 GB / s | I-736 GB / s | I-504 GB / s | I-504 GB / s |
| I-TBP | 450W | 320W | 285W | 200W |
| Intengo (MSRP / FE) | $1599 US / 1949 EU | $1199 US / 1469 EU | US $ 799 US | US $ 599 US |
| Intengo (Okwamanje) | $1599 US / 1859 EU | $1199 US / 1399 EU | US $ 799 US | US $ 599 US |
| Qalisa (Ukutholakala) | I-XNUMi-Okthoba i-12 | I-16th ngo-November 2022 | I-XNUM ngo-Januwari 5 | I-13th Ngo-Ephreli 2023 |
Ibhalwe ngu Hassan MukhtabaIsiqephu Sokuqala
