Radeon RX 5600 XT with new vBIOS boost shows impressive new benchmark scores

onetheycallEric

Posts: 225   +47
Staff
Recap: This battle between AMD and Nvidia essentially started at CES 2020, shortly after AMD announced its RX 5600-series. EVGA, presumably with Nvidia's blessing, then unveiled what seemed a preemptive response in the form of RTX 2060 KO graphics cards, starting at the same $279 price as AMD's RX 5600 XT.

Then, we learned that AMD pushed out a new vBIOS to its AIB partners as a response to Nvidia's price cuts on the RTX 2060 in what's been a back and forth volley between the two companies at the $300 price point.

The new vBIOS ratchets up the TBP (Typical Board Power) to 160W, a 10W increase from the original 150W TBP. This in turn affords AMD's board partners a higher margin for core and memory clocks. Which brings us to Sapphire, one of AMD's premiere AIB partners.

As we previously reported, Sapphire's RX 5600 XT Pulse is now boasting increased base and boost clocks of 1,615 MHz and 1,750 MHz, up from the previous 1,560 MHz and 1,620 MHz. The memory clock is now running at 14Gpbs effective, as opposed to 12Gpbs.

Now, Twitter user @TUM_APISAK has sniffed out a leaked benchmark showing what kind of performance uplift the new vBIOS can create.

The benchmarks are, in order: Time Spy, Fire Strike, Fire Strike Extreme, and Fire Strike Ultra. Doing the math between the new and old scores, the new scores net an impressive 10% - 11% uplift in performance, depending on the benchmark.

As with all synthetic benchmarks, these don't necessarily indicate what users can expect in the real world. However, it's certainly a nice preview of what we can expect from the RX 5600 XT. AMD and Nvidia will likely continue to insist that this was the plan all along, rather than either company admit to being one-upped by the other. However, the more pertinent question is: are there any more cards left to play, or have AMD and Nvidia showed their full hand?

Permalink to story.

 
It is probably within 5 percent of a 5700 now going by those benchmarks. Although you can get one of those with a decent cooler for 10 percent more money, or $310. I would be tempted to go a little higher for the extra memory and bandwidth even so.

The GPU market has been carpet bombed with models which isn't necessarily a bad thing, it just gives a lot of options at prices all close together. Best to be on the look out for deals.
 
The next thing Nvidia should do is keep cutting prices of its cards to meet the AMD cards at their launch MSRP.

Wouldn't hurt to cut the 2080Ti a $100 - or to meet whatever their competitor will start at.
 
With 5600 being pushed so close to 5700, wouldn't 5700 still be the better chip overall and thus the better buy? Especially for longevity and stability. Slap a turbo on a 4 cyl. and it might keep up with a 6 cyl. but the 4 cyl. is on the verge of blowing up in order to do so. I think AMD would be better serve to just let 5600 be and drop the price on both instead.
 
With 5600 being pushed so close to 5700, wouldn't 5700 still be the better chip overall and thus the better buy? Especially for longevity and stability. Slap a turbo on a 4 cyl. and it might keep up with a 6 cyl. but the 4 cyl. is on the verge of blowing up in order to do so. I think AMD would be better serve to just let 5600 be and drop the price on both instead.
They are the same chip. The only difference is the ram. Your analogy is wrong.
 
With 5600 being pushed so close to 5700, wouldn't 5700 still be the better chip overall and thus the better buy? Especially for longevity and stability. Slap a turbo on a 4 cyl. and it might keep up with a 6 cyl. but the 4 cyl. is on the verge of blowing up in order to do so. I think AMD would be better serve to just let 5600 be and drop the price on both instead.

That depends on whether AMD was even running their "4 cylinder" to it's full potential to begin with. Given that we are only seeing a 10w increase for that much extra clock speed, it appears to me that AMD is still running close to the sweet spot on these cards and they likely have even more room to OC. If the 5700 / 5700 XT safely overclocks to 2.1 GHz, these chips at 1.7 may very well still be conservatively clocked.

It would be easy to tell if AMD were pushing the cards too far this late in the development cycle. Too much voltage would be required and the coolers (designed for the original 150w TDP) would have trouble dissipating the heat.

I should point out that the RTX 2060 also has a 160w TDP.
 
They are the same chip. The only difference is the ram. Your analogy is wrong.
Same chips, but different quality levels, the cause for the analogy. Granted AMD may have allocated a certain amount to each class where a lower class chip may actually be on the cusp of higher quality and will thus be part of the lottery class of chips. For the most part, it's a fair analogy.
 
If nvidia could they would do the same.
From my experience if you push Nvidia cards to hard they just stop working.
better buy the founder cards for better quality components.
 
Same chips, but different quality levels, the cause for the analogy. Granted AMD may have allocated a certain amount to each class where a lower class chip may actually be on the cusp of higher quality and will thus be part of the lottery class of chips. For the most part, it's a fair analogy.
Like the R9 290X was? or one of the older model I can't remember which one. Where Micron and Hynix and Samsung greatly affected how you can mod the vBIOS.

If nvidia could they would do the same.
From my experience if you push Nvidia cards to hard they just stop working.
better buy the founder cards for better quality components.
Nvidia doesn't have to they are leading ATM, but this year we will see what Intel and new AMD cards will bring. This RX line doesn't give me too much hope for competition in HEDT.
Also confusing to have different naming system in the last couple years.
 
If you buy a 5700, you could always flash it to a 5700 xt, getting a 10%ish? boost 1910 core etc
 
Last edited:
Back