• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD Navi 23 ‘NVIDIA Killer’ GPU Rumored to Support Hardware Ray Tracing, Coming Next Year

Status
Not open for further replies.
g67575 said:
Why do some ppl want RDNA 2 graphics cards to have HBM2 memory? Memory bandwidth is likely to be the least of the RX 6000 series' problems.

Using HBM2 rather than GDDR6 will just add to the cost too and require more power that could be used for other things like higher clocks or more CUs. Probably better to use a higher spec memory bus to increase memory bandwidth.
Click to expand...

HBM2 is lower power draw, so you’ve got that bit backwards.

As for why, I think most of it stems from the rumour that the bus is 256bit even on big dog Navi and the assumption (Rightly or wrongly) is that it will be limiting on a card that is supposed to compete with the 3080 and GDDR6X on a 384 bit bus.

Had the rumours continued to point towards a 512bit bus and GDDR6 I guess there’d be less people hoping for HBM2.
 
g67575 said:
Why do some ppl want RDNA 2 graphics cards to have HBM2 memory? Memory bandwidth is likely to be the least of the RX 6000 series' problems.

Using HBM2 rather than GDDR6 will just add to the cost too and require more power that could be used for other things like higher clocks or more CUs. Probably better to use a higher spec memory bus to increase memory bandwidth.
Click to expand...

I remember when certain people here were convinced, probably because AMD had hyped it a bit, that HBM was the future and was going to destroy nVidia who would still be using GDDR, etc. and bury GDDR. Got some quite pitiful and child like posts aimed at me for saying GDDR would not be obsolete any time soon, would be able to keep pace via moving to small processes, etc. and that HBM isn't a superior replacement for gaming cards...

HBM has some advantages but the patterns of memory access while rendering a video game don't necessarily allow you to leverage those advantages and in some cases can hit disadvantages over GDDR - like with HDD/SSD storage there is more to the story than the straight up headline sustained transfer rates and latencies at a given load level.

james.miller said:
Yeah, what was i thinking
Click to expand...

Wish nVidia would go back to those Pascal/Turing style shroud/heatsink designs - some of them are a work of art when you notice the little details that have been put into them - albeit they are a pain in the rear if you want to remove them for any reason. The Ampere one looks a distinct step backwards in comparison.
 
james.miller said:
Yeah, what was i thinking :p

RX_6xxx_b.jpg

QcYBDpxh.jpg




solid reasoning.

Linus said the 2080ti prices were inflated and he didnt recommend the 2070 based on the complete lack of RTX benchmarks available, and LTT ripped Tom's Hardware's 'opinion piece' on why RTX cards were worth the money. But no, he's a hypocrite right?
Click to expand...

ouch lol, AMD seems be behind again - copies a NVIDIA’s cooler not knowing that's the old design and the one is better looking and much better cooling
 
I wouldn't call it an exact copy




Other then the location of the branding "RADEON" it's a bit of reach to think it's copying Nvidia



The above is from Sapphire and this is what any typical 3 axial hsf looks like.

----------------
HBM saves on both space and saved energy consumption over GDDR. The down side was capacity and cost. However, with HBM2e it address capacity and speed.

Samsung’s new HBM2E (sold under the Flashbolt brand name, versus the older Aquabolt and Flarebolt brands), offers 33% better performance over HBM2 thanks to doubling the density to 16 gigabits per die. What’s changed is these devices are based on eight 16-Gbit memory dies interconnected using through-silicon vias (TSVs) in an 8-Hi stack configuration. That means a single package is capable of 410 GB/s memory bandwidth and 16GB of capacity. With double per-die pin capacity, DRAM transfer speeds per pin can reach 3.2 Gbps.


With HBM2E, AMD, one of the co-developers of HBM, is turning the same levers again. “The only bits added to the interface were to increase addressability, but it’s the same interface, it just runs at a higher interface of 3.2 gigatransfers per second,” Macri said.


From 2 to 2E
The change from HBM2 to 2E is not revolutionary. It is pretty much a speeds and feeds update, but that’s more than enough for now, said Samsung.


“The key motivation in market outreach with HBM2E is higher capacity, and HBM3 will have even higher bandwidth and more capacity,” said Tien Shiah, senior manager for memory at Samsung. “


Then there are HBM3+ and HBM4, reportedly to be released between 2022 and 2024, with more stacking and higher capacity. HBM3+ is supposed to offer 4 TB/s throughput and 1024 GB addressable memory per socket.
Click to expand...

Will AMD use HBM2e? Who knows...we have to wait until the 6900 is revealed.
 
Last edited:
g67575 said:
Why do some ppl want RDNA 2 graphics cards to have HBM2 memory? Memory bandwidth is likely to be the least of the RX 6000 series' problems.

Using HBM2 rather than GDDR6 will just add to the cost too and require more power that could be used for other things like higher clocks or more CUs. Probably better to use a higher spec memory bus to increase memory bandwidth.
Click to expand...

I believe folks are speculating based on fortnite reveal.. though I would like to be pragmatic about it and suggest that those renders were creative works... only meant to reveal key features.. and probably not-to-scale
 
Last edited:
Well I like the red piping on these cards personally, not that it matters a jot because it'll be hidden in my case under my desk for months on end. That Jayz fella clearly isn't a great fan of AMD.

Will it run MFS2020 at 4k60 ultra and cost less than the 3080 is what bothers me.
 
I'm liking the look of the new cards. If I wasn't skint I'd buy one, assuming the performance to price was good.

As far as Techtubers go I think my favourite channel is Hardware Unboxed. Good thorough reviews, buy a lot of their own components to test and happily call out dodgy practices.
 
Trick said:
nice one! What are your voltage settings for p5 p6 and p7?

I might try going a little lower on mine, but I’m wondering if it’ll boost up to 1630mhz still with less power?
Click to expand...

State 7 1070 - 1645
State 6 1020 - 1538
State 5 970 - 1474

This is on vega 64 bios with 50% power limit and mem on 1050 and 1000mv

mine doesnt boost to anywhere close to 1645 stays around 1600mhz.
For more i have to pump a lot more juice, then fans ramp up too much esp on a hot day.
 
Status
Not open for further replies.
Back
Top Bottom