Rumors about Nvidia’s next-gen graphics cards from the Geforce RTX 5000 series (“Blackwell”) have been in the headlines for weeks. PCGH tries to classify these seriously based on the current situation. Once again, the Twitter user @kopite7kimi, who is known for such pre-releases, wants to know which graphics processors will be used on the next-generation Geforce graphics cards. Accordingly, a direct successor to the AD104 is not expected to be planned.
GB202, GB203, GB205, GB206 and GB207 for Blackwell
In the meantime, the user “Panzerlied” from the Chiphell forum, which is known for such information, has also confirmed that he has information that the GB202, GB203, GB205, GB206 and GB207 graphics processors for the upcoming Geforce RTX 5000 series graphics cards (“Blackwell “), which is currently also traded as “Ada Lovelace Next”, confirm. Whether the envisaged generation change will take place in 2024 or only in 2025 is still open.
Geforce RTX 5090 and 5080 are probably the first
The graphics processors AD102, AD103, AD104, AD106 and AD107 (“Ada Lovelace”) are expected to be followed by the new GPUs GB202 and GB203 (“David Blackwell”). Numerous other expansion stages with an even higher identifier are also expected here, which are intended for the Geforce RTX 5070 (Ti) and the smaller model series. Nvidia should remain true to its scheme.
GDDR7 with 36 GiBit/s and PCIe 5.0 x16 are considered set
In addition to a connection via PCIe 5.0 x16, the new GDDR7 graphics memory should be used at least for the Geforce RTX 5090 and Geforce RTX 5080. This should reach up to 36 gigabits per second and rely on three-stage pulse amplitude modulation (“PAM-3”) for signal transmission. It can be assumed that energy efficiency in this sub-area has improved by around 25 percent. GDDR6X, on the other hand, already uses PAM-4.
The effective memory bandwidth would also benefit significantly from the fast GDDR7 memory with up to 36 GiBit/s:
- 128-bit @ 36Gbps: 576 GiBytes/s
- 192-bit @ 36Gbps: 846 GiBytes/s
- 256-bit @ 36Gbps: 1,152 GiBytes/s
- 320-bit @ 36Gbps: 1,440 GiBytes/s
- 384-bit @ 36Gbps: 1,728 GiBytes/s
- 512-bit @ 36Gbps: 2,304 GiBytes/s
As can be seen from several entries in the Chiphell forum, Nvidia could also take a two-pronged approach:
- GB202: GDDR6X (“PAM-4”) with 512-bit
- GB203: GDDR6X (“PAM-4”) with 384-bit
- GB205: GDDR7 (“PAM-3”) with 256-bit
- GB206: GDDR7 (“PAM-3”) with 192-bit
- GB207: GDDR7 (“PAM-3”) with 128-bit
But even this is usually still pure speculation, which is based on information that has been leaked so far, but is not yet to be regarded as absolutely binding. Nvidia itself has not said a word about the next generation of its Geforce graphics cards.
Performance forecasting is almost impossible
The rumor mill says the upcoming top model, the Geforce RTX 5090, will have up to 2.5 times the performance of an already very fast Geforce RTX 4090. For the time being, however, great caution is advisable, because it is currently still completely unclear whether DLSS 2 or even DLSS 3 with frame generation will be included in the equation. Performance forecasts are currently almost impossible.
Geforce RTX 5090 with up to 18,432 CUDA cores
The specifications assigned to the Geforce RTX 5090 are also purely speculative. According to this, it should go up from 128 streaming multiprocessors and 16,384 shader units to 144 streaming multiprocessors and 18,432 CUDA cores. In addition, the L2 cache should be increased from 72 MiByte to 96 MiByte and the GPU clock should increase to more than 3 GHz. Videocardz has already compared the graphics processors of the three generations.
The PCGH editorial team will report on this if correspondingly valid information about the Nvidia Geforce RTX 5000 is available.
Source: Videocardz, Chiphell, kopite7kimi via Twitter
