New Posts  All Forums:

Posts by tpi2007

I wonder if people were able to play the same games on Vega FE and RX Vega:https://twitter.com/Radeon/status/889187135286370304Notice the Radeon RX bracelets on top of the rig.Two years late would be impossible, since the Fury X was launched two years ago (Fury X - 24 June 2015 and Fury - 14 July 2015).But there have been numerous articles where Vega is mentioned since at least March 2016; the first Vega related article on TPU is from Mar 15th, 2016. See below for a...
Here is one thing that definitely doesn't compute with those people saying that Vega FE is not RX Vega and that Vega FE is not a gaming card (even though Raja said that Vega FE supports the RX driver and they have a paragraph in the Vega FE page instructing people who "just want to game" to switch control panels):This is what PDXLAN is about:http://www.pdxlan.net/home/What does AMD do besides bringing RX Vega? They also bring Vega FE:https://twitter.com/Radeon"Marketing...
They revised the liquid cooling version's TDP down to 350w at the last minute. Really at the last minute, as in the same week that the air cooled version was released they were still publicly saying the liquid cooled version had a 375w TDP. 375w would be at the certified limit of the PCIe + 2x 8 pin connectors, so it's understandable why they did it, to prevent another potential Polaris 10 PCIe certification problem, where certain samples would go over the limit, but it...
There were other options to consider if you have another HDD / SSD where you were going to move the whole game to: move some other big game onto that drive, thus freeing up the space needed to complete the update on the Doom game. Unless we're talking about an 80 GB HDD / SSD from which there is nothing else to move.As to the general issue, Steam lacks granularity and freedom when it comes to updates. If someone just wants to play single-player, why force updates? That...
He said in the first tweet that it was running at 1.25v: https://twitter.com/HardwareUnboxed/status/888225519141093376I wonder how HardOCP's 7900X is after the article they made. Kyle does note this:https://www.hardocp.com/article/2017/07/19/overclocking_intel_core_i97900x_xseries_processorFrom these two experiences, on a 7800X and a 7900X, it's probably best to stay safe and avoid going higher than 1.2v, which in itself seems sufficient and even the useful max for very...
According to the very same person, Vega X2 is a 7nm product.Thus he only forgot one convenient detail in the post above: it'll be out in full force by X-mas alright. Of 2018, that is.Source:Also, please discard this completely not official information that says that it's Navi that will be on 7nm and not Vega:
What we had before were 3DMark 11 results (vs these Fire Strike Extreme results) with the card at 1,630 Mhz, so making comparisons is impossible as the variables are all over the place.But being better than the 1080 is what most people are expecting anyway, so this result does fall in line with those expectations. We'll see at what power cost those clocks were achieved. Nobody thought that the FX-9590 and its 220w TDP was a good idea except for the first ever symbolic 5...
Both the CPU and the GPU were clocked to the moon in that result.Ryzen 7 1800X at 4225 Mhz - something you don't see everyday and presumably RX Vega clocked at 1,756 MHz, which probably won't do well in the power consumption tests from what we've seen so far from Vega FE, even with undervolting.On the other side you have the same Ryzen but at 4 Ghz, that's a 5%+ difference in CPU clocks and then you have the 1080 Ti at 1,886 MHz, so it's not maxed out; there's at least...
Of course a smaller die can be a winning strategy, but only when you look at a specific product and not talk in abstract - the GTX 1060 proves that because Nvidia figured out how balance performance and clocks in their architecture and came up with a 200 mm² die and a 120w TDP, which is both smaller and more power efficient than AMD's chip with which it competes. If AMD starts out by losing (and the GTX 1060 is much more popular than the RX 480), they might as well have...
They wanted to cut costs as much as they could, so in order to get as many dies per wafer as possible, they designed the smallest chip possible, then clocked it beyond its operating efficiency range. The same with Fiji (although in that case they had a good excuse, as the chip couldn't be any bigger); at 900 Mhz average it's a 175w chip, but at 1050 Mhz it's a 275w chip. Polaris 10 and 20 would probably have been a much better product, including in the longer term, as we...
New Posts  All Forums: