Soldato
- Joined
- 22 Nov 2018
- Posts
- 2,840
Actually you're right. I've just been looking at benchmarks and that Navi card only has to beat the Vega 56 by 10% and it will equal a 1080 so $250 is realistic.
-
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 Zen 2 (Ryzen 3000) - *** NO COMPETITOR HINTING ***
- Thread starter Jamin280672
- Start date
Actually you're right. I've just been looking at benchmarks and that Navi card only has to beat the Vega 56 by 10% and it will equal a 1080 so $250 is realistic.
AMD Ryzen 3000 Desktop CPUs, Radeon Navi GPUs and X570 Motherboards With PCIe Gen 4.0 Allegedly Launching on 7th July
https://wccftech.com/amd-ryzen-3000-cpus-x570-motherboards-and-radeon-navi-gpus-7nm-launch-rumor/
https://wccftech.com/amd-ryzen-3000-cpus-x570-motherboards-and-radeon-navi-gpus-7nm-launch-rumor/
Soldato
- Joined
- 26 Oct 2013
- Posts
- 4,132
- Location
- Scotland
Remember though, the Radeon VII was never originally intended to be a gaming card and was released as a kneejerk reaction to a lot of defective die or not enough sales in the professional space so isn't really going to be indicative of AMD's next generation architecture being designed for gaming.
1080 (a card released in 2016) performance for $250 is where we SHOULD be aiming 3 years later, though I know it has been slow since the release of Pascal cards compared to generations previous.
Because Radeon VII is a PR stunt made possible through Nvidia's snafu with RTX, not an actual and proper gaming product.
Nvidia come in with the RTX 2080 which features underpowered technology utilised by very little software at the expense of boosting traditional raster performance, all for a price tag that takes the mickey. Vega 20 was never intended to be a gaming GPU, but through brute force it can match the RTX 2080 in raster performance, so there's suddenly the opportunity to repurpose failed MI50 packages as a gaming product and suddenly release something that is on par with a top-end Nvidia product that's only a few months old.
Now, you could argue that since those MI50 packages were going in the trash, AMD could easily have charged $500 for Radeon VII since they were going to lose money anyway. This is true, but doing so would risk skewing perceptions of Navi's pricing later on: if a gaming card with 16GB of expensive HBM2 memory can match the RTX 2080 at $500, why would a card that can only match the RTX 2070 with half the amount of cheap GDDR6 cost as much as $250-300? Navi is a gaming-first architecture designed to be 7nm from its inception. Small, lean, cheap and fast (ish), that is where the potential to undercut Nvidia comes from; keep the HBM2 monster at Nvidia prices, but then release the lean and focussed product at a sensible price.
I think 2019 is the year where AMD attempt to bring some kind of sanity to CPU and GPU prices, but Radeon VII is not the product to do that. Ryzen 3000 will shift perceptions of just how much grunt and cores you can get on the desktop for a sensible amount of money, EPYC will turn major heads in what servers and datacenters can be equipped with for sensible money, Navi will shift perceptions on how much 2 year old gaming performance should cost.
And I think the PR stunt that is Radeon VII will be replaced by a "Big" Navi in early 2020 - same performance but much less power and noise for about $450, although I'd be surprised to see 16GB of GDDR6 (Radeon VII has set that precedent now, unfortunately).
Soldato
- Joined
- 22 Nov 2018
- Posts
- 2,840
You would think so but we don't yet have 980 ti (a card released in 2015) performance for $250.
It has been taking longer than usual for that level of performance to become affordable.
Soldato
- Joined
- 26 Oct 2013
- Posts
- 4,132
- Location
- Scotland
That's the power of a lack of competition sadly. I hope the market picks up pace again within the next year or 2.
7th Jusly would make sense being 7nm I guess. I was really holing to be a May/June launch though, I'm so ready for a new build now.
Given PCIe 4, I think it was always going to be optimistic that the CPUs would launch before the X570 motherboards, and that leaked Gigabyte slide from way back when pushed "Matisse" as an all-PCIe 4 product with X570 launching June - the first half at least is confirmed. There are suggestions that X570 itself is a bit behind, so I'm expecting now Computex will announce X570, show some demo boards, and the Ryzen 3, 5 and 7 SKUs confirmed too (the 9 SKU will be held back*). Then we see if 7/7 gives us the actual launch. Even taking the alleged re-taping of Navi out of the equation, I think July is too early. I had it in my head for some reason that September was more realistic for Navi.
It remains to be seen though what will be announced on May 1st; surely AMD must do something big for their 50th anniversary. I'd love a full-on 3850X demo
"you saw our engineering 8 core like-for-like test against the 9900K, now here's Ryzen fully unleashed (and we've not even done Threadripper yet)".* I still believe the Ryzen 9 3800X will be the official 16-core SKU, but AMD will ride the hype train for the anniversary 3850X. It's the same CPU but the limited run is binned and confirmed to hit 5.1GHz boost. Do 50,000 units of these for a nice premium, then after they sell out confirm the 3800X as the ongoing Ryzen 9 SKU, and if you win the silicon lottery you can still push up to 3850X equivalent. Kinda like the 8700K/8086K but in reverse.
Last edited:
Hmm 7th July is later than I'd hoped, especially given they'll likely start with the halo products and the ones I'd be more interested in might not appear for an extra couple of months after that. At least if Epyc 2 is launching in June we'll know much more about IPC improvements then.
For me (on an Asus Prime X470) - a flagship with PCIe 4 is totally happening.
Chip wise, the one with the fastest boost clocks. Bonus points if it's the 3850X 16/32 @ 5.1ghz.
This is why it is SOOOOO hard for me to justify a purchase. The only thing that has really moved on are the prices... upwards.
I am extremely excited about the CPU space though. Real gains are being made here in VFM.
Nothing is confirmed. All we've seen is that 8c/16t engineering demo, Lisa Su saying "yes there's space for another chiplet, you'd expect us to add more cores" and some 12c/24t 3.7GHz benchmark going around.
The die shrink and double the amount of transistors getting to a stop soon.
more expensive to make means prices will go up.
No idea how they are solving the 3nm sizes as walls are thin then
Soldato
- Joined
- 16 May 2007
- Posts
- 3,221
I bought Ryzen 1 waiting for significant gains before updating to the later Ryzen versions. As soon as it hits high 4’s or 5 Ghz then I will upgrade. The 3700x with 12 cores at around 5.0Ghz based on rumours looks like my next upgrade if this holds true.
I only upgrade if there is a decent jump in gaming performance from my 1st gen ryzen