Jump to content

Does running a 24gb GPU on pcle5.0 x8 slot impact the performance vs running it at pcle5.0 x16

Bihaidi

Does running a 24gb GPU on pcle5.0 x8 slot impact the performance vs running it at pcle5.0 x16, I am currently trying to choose a mobo for my build and looking at the ASUS ROG STRIX X670E-E GAMING WIFI, that board has 4 m,2 slots (3 are pcle5.0 running from the CPU, the last one is a pcle4.0 from the chipset) now then if I were to populate all 3 m.2 keys with pcle5.0 slots that would leave about 12/24 lanes leftover meaning 8 usable lanes at pcle5.0 for the gpu, that got me curious about how the number of lanes would impacts the performance for context the pcle5.0 bandwidth is 63 gb at 16 lanes and 32 gb at 8 lanes, and since most modern GPU'S dont really go above 24gb it should technically run the same, but I am fairly new to building computers so don't really know and would appreciate if anyone could kindly leave a comment below with the answer.

Link to comment
Share on other sites

Link to post
Share on other sites

The only PCIe 5.0 gpu out there is the Nvidia H100 PCIe with 80GB HBM2e.

Interim 15 T200 OKF("F" intel processors are specifically archituctured for gaming) maybe upgrad to 13'900 | Peeralight cpu fan | Stryx Z690-A Wife(which is branded by ASUS and it's ROG label) | Thermotake 16x 8x2GO SODINM 2400mjz cl22 (2 of them with the mood lighting) | 980 EVO 1TB m.2 ssd card + Kensington 2TB SATA nvme + WD BLACK PRO ULTRA MAX 4TB GAMING DESTROYER HHD | Echa etc 3060 duel fan dissipator 12 GBi and Azrock with the radian 550 XT Tiachi | NEXT H510 Vit Klar Svart | Seasonice 600watts voeding(rated for 100.000 hours, running since 2010, ballpark estimate 8 hours a day which should make it good for 34 years) | Nocturna case fans | 0LED Duel moniter

 

New build in progress: Ryen™ 8 7700x3D with a copper pipe fan | Z60e-A | Kingstron RENEGATE 16x2 Go hyenix | Phantek 2 the thar mesh in front | lead lex black label psu + AsiaHorse białe/białe | 1080 Pro 8TB 15800MB/S NvMe(for gaming this increase fps and charging time, cooled by a M.2 slot with coolblock and additional thermopad) and faster 4000GB HHD | MAI GeForce GTX 2070 Ti and RTX 6800 | Corshair psu

Link to comment
Share on other sites

Link to post
Share on other sites

I.e. your GPU will be running at PCIe4 x8.

Check if someone did performance comparison at PCIe4 x8 / x16 with the GPU you're after.

VRAM amount doesn't matter.

F@H
Desktop: i9-13900K, ASUS Z790-E, 64GB DDR5-6000 CL36, RTX3080, 2TB MP600 Pro XT, 2TB SX8200Pro, 2x16TB Ironwolf RAID0, Corsair HX1200, Antec Vortex 360 AIO, Thermaltake Versa H25 TG, Samsung 4K curved 49" TV, 23" secondary, Mountain Everest Max

Mobile SFF rig: i9-9900K, Noctua NH-L9i, Asrock Z390 Phantom ITX-AC, 32GB, GTX1070, 2x1TB SX8200Pro RAID0, 2x5TB 2.5" HDD RAID0, Athena 500W Flex (Noctua fan), Custom 4.7l 3D printed case

 

Asus Zenbook UM325UA, Ryzen 7 5700u, 16GB, 1TB, OLED

 

GPD Win 2

Link to comment
Share on other sites

Link to post
Share on other sites

16 minutes ago, Bihaidi said:

bandwidth is 63 gb at 16 lanes and 32 gb at 8 lanes, and since most modern GPU'S dont really go above 24gb it should technically run the same

This is not how it works. Even if they had 1TB of vram it wouldn't automatically mean they could saturate pcie 5 16x bandwidth. As mentioned there aren't even any consumer GPUs using PCIE 5 yet, however it's unlikely any current GPU would saturate even a 4.0 x8 slot. As you can see there is barely any difference between 4.0 16x and 3.0 16x which is equivalent to 4.0 8x:

 

f1-2022-3840-2160.png

 

In fact 2.0 is still fine, which has the equivalent bandwidth of 5.0 2x.

Don't ask to ask, just ask... please 🤨

sudo chmod -R 000 /*

Link to comment
Share on other sites

Link to post
Share on other sites

45 minutes ago, Sauron said:

This is not how it works. Even if they had 1TB of vram it wouldn't automatically mean they could saturate pcie 5 16x bandwidth. As mentioned there aren't even any consumer GPUs using PCIE 5 yet, however it's unlikely any current GPU would saturate even a 4.0 x8 slot. As you can see there is barely any difference between 4.0 16x and 3.0 16x which is equivalent to 4.0 8x:

 

f1-2022-3840-2160.png

 

In fact 2.0 is still fine, which has the equivalent bandwidth of 5.0 2x.

It does vary depending on the game but a 3.0x16 link is basically fine for all

Link to comment
Share on other sites

Link to post
Share on other sites

3 hours ago, Kilrah said:

I.e. your GPU will be running at PCIe4 x8.

Check if someone did performance comparison at PCIe4 x8 / x16 with the GPU you're after.

VRAM amount doesn't matter.

I've seen cards with insufficient VRAM getting hit harder by PCI bandwidth.

AMD 7950x / Asus Strix B650E / 64GB @ 6000c30 / 2TB Samsung 980 Pro Heatsink 4.0x4 / 7.68TB Samsung PM9A3 / 3.84TB Samsung PM983 / 44TB Synology 1522+ / MSI Gaming Trio 4090 / EVGA G6 1000w /Thermaltake View71 / LG C1 48in OLED

Custom water loop EK Vector AM4, D5 pump, Coolstream 420 radiator

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×