AMD’s brand new high-performance GPU architecture – Vega – removes many traditional constraints from Gaming, Virtual Reality, Professional Design and Machine Intelligence arenas. AMD’s Vega architecture is one Nvidia will be closely looking at. The new architecture from AMD enables new possibilities in PC gaming, professional design, and machine intelligence that traditional GPU architectures have not been able to address effectively. Vega’s memory subsystem enables GPUs to address very large data sets spread across a mix of memory types.
With 2x peak FP16 throughput/performance per clock, High Bandwidth cache, 2x Bandwidth per pin, 8x capacity per stack, 512TB virtual address space, Next Gen Compute Engine, Pixel Engine, Compute Unit optimized for higher clock speeds, rapid packed math and primitive shaders, AMD Vega architecture is expected to deliver rock solid performance.
AMD Vega Dual GPU liquid cooled graphics card spotted
Linux drivers have always been the treasure trove of information about AMD’s products. AMD released a new brand new Linux driver yesterday. And a liquid cooled dual GPU Radeon RX Vega graphics card has been spotted in the Linux driver. A Redditor has discovered interesting lines of code that have been added to AMD’s Linux open-source driver stack. Two more PCI devices IDs for the Vega 10 family have been added – 0x6864 and 0x6868.
table->Tliquid1Limit = cpu_to_le16(tdp_table->usTemperatureLimitLiquid1);
table->Tliquid2Limit = cpu_to_le16(tdp_table->usTemperatureLimitLiquid2);
As for the dual GPU board, the redditor found some more valuable pieces of code –
table->FanGainPlx = hwmgr->thermal_controller. advanceFanControlParameters.usFanGainPlx;
table->TplxLimit = cpu_to_le16(tdp_table->usTemperatureLimitPlx);
This code is probably specific to PLX chip, found in Vega 10 code stack. PLX is the company that manufactures PCI Express Switch. The number of PCI Express lanes is limited by what the CPU provides. A PCIe switch allows more slots on the motherboard to be wired with lanes, let’s say 16-lane CPU. A switch allows for 32 active lanes at the same time.
PLX chips on AMD cards are used in order to bridge two GPU’s on a single board by multiplexing the PCI-Express slot on the motherboard to both GPU’s. The presence of PLX thermal management code strongly indicates that a dual GPU vega board is on the way.
This is all conjecture on my part, so please take it all with a pinch of salt. It is still possible (though I believe unlikely) that these are remnants from Pro-Duo/Fiji driver code.
Don’t forget to check: Top 10 Upcoming PC Games In 2017