This site may earn chapter commissions from the links on this page. Terms of use.

AMD Radeon RX Vega 56

When Nvidia or AMD launch a new GPU, there's a typical rollout pattern. The first cards out the door are reference designs, based on a packet Nvidia and AMD provide. After, companies similar MSI, Gigabyte, EVGA, Asus, and PowerColor debut their own custom designs. These custom boards are mostly clocked higher, have better, quieter coolers, or may be built on a custom smaller PCB. This time around, notwithstanding, AMD'southward lath partners are reportedly having major problem sourcing GPUs.

That's the discussion from Tom'due south Hardware, which spoke to a number of companies. XFX and Sapphire accept cards in the works, but no business firm release date. PowerColor has a design only has yet to receive the DRAM it needs. Asus has pushed back the launch date on its ain cards, Gigabyte hasn't firmly committed, and MSI is apparently skipping the party altogether.

THG writes that the problem is twofold. First, the quality of AMD'south Vega GPUs has been variable enough that OEMs don't experience comfortable establishing an overclock frequency they can guarantee and ship. Second, there have been temperature reporting discrepancies between what the board reports and what AIBs are measuring.

MoldedvsUnmolded

Molded vs. unmolded GPU packaging.

Finally, at that place are mounting differences that'south making information technology hard to mass produce Vega. Dissimilar previous GPUs, non every Vega card uses the same mounting. There are three variants of Vega in-market: Ane with HBM that's the aforementioned height every bit the GPU (molded, Samsung HBM2), one with HBM that's 40 micrometers lower than the GPU (unmolded, Samsung HBM2) and apparently a third variant with Hynix memory.

Was HBM a Mistake?

I'k not going to pretend to have a secret source inside AMD on this one, but information technology'south difficult to look at Vega and not wonder if HBM was a fundamentally bad phone call. The visitor'due south molding issues are directly related to the height of the HBM2 memory. Meanwhile, it took AMD over two years to launch Vega. That'due south the longest gap AMD or ATI has ever gone between high-cease GPU refreshes, which previously took 12 – 16 months at most. HBM2'south rollout has been slower than anticipated in general, specially the highest speed memory.

Nosotros don't know for certain that HBM is the culprit here, but it certainly seems like the about obvious place to wait. Because there's always a considerable lag betwixt when a GPU design projection kicks off and when the final product tapes out, it's entirely possible that AMD was also far forth in the pattern process to get-go over when HBM2's growing pains became clear.

The large point hither will be whether AMD's futurity GPUs are based on GDDR6. If HBM2 was the culprit for the issues Vega seems to have, AMD won't keep using information technology. AIBs are yet expected to bring custom Vega cards to market in general, but those samples may non be in-market until closer to the holiday season.