New Zen microarchitecture details

Page 198 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Magic Hate Ball

Senior member
Feb 2, 2017
290
250
96
Why not use original source, then?
https://www.youtube.com/watch?v=4DEfj2MRLtA from 47:00 or about that.
On one side they had 6900k that did identically to Ryzen rig and on the other they had 6700k that lagged. Infinity fabric was never mentioned.

https://youtu.be/vMfNz2SXVLk?t=2m48s

This is where the info came from regarding Infinity Fabric.

Linus was at the CES 2017 show in their booth and had spoken with the AMD reps regarding the longer delay between the 6900K and the streaming feed versus the Ryzen's feed.

To further clarify, this is not about frames being dropped or stuttering video. This is only in regards to a delay in feed processing.
 

Agent-47

Senior member
Jan 17, 2017
290
249
76
Why not use original source, then?
https://www.youtube.com/watch?v=4DEfj2MRLtA from 47:00 or about that.
On one side they had 6900k that did identically to Ryzen rig and on the other they had 6700k that lagged. Infinity fabric was never mentioned.
??
Why choose to ignore reliable evidence that does not agree with you and go for something that does not even address the question we are trying to answer?

I don't think we are talking about the same thing here. in new horizon they were showing frame drops while streaming (not surprising as its 8c vs 4c, one might add) , while the Linus vedio shows stream delay. Not the same thing I am afraid
 

lolfail9001

Golden Member
Sep 9, 2016
1,056
353
96
Linus was at the CES 2017 show in their booth and had spoken with the AMD reps regarding the longer delay between the 6900K and the streaming feed versus the Ryzen's feed.
Well, i see now, except that i could not quantify delay on video because replays are de-synced. Eh.
I don't think we are talking about the same thing here
Yeah, i see now, except that i do not buy it just yet.
 

DrMrLordX

Lifer
Apr 27, 2000
21,583
10,785
136
1)
An 8c16t Ryzen will compete with the 8c16t Intel equivalent, not a $350 7700k. Your entire post is invalid.

I think AMD would market the top-end Ryzen against the i7-7700k. Basically a "look moar cores" argument since they know they can't match both the IPC and clockspeed of a chip like that, nor do they want to shoot for the astronomical prices or platform features of X99 or X299.

They'll bench vs the 6900k as a pissing contest but when it comes to pricing/marketing, the primary target is Z270/7700k.
 

Agent-47

Senior member
Jan 17, 2017
290
249
76
I think AMD would market the top-end Ryzen against the i7-7700k. Basically a "look moar cores" argument since they know they can't match both the IPC and clockspeed of a chip like that, nor do they want to shoot for the astronomical prices or platform features of X99 or X299.

They'll bench vs the 6900k as a pissing contest but when it comes to pricing/marketing, the primary target is Z270/7700k.

If a 8c16t will be having a pissung contest with 6900k, why would they price it against 7700k? They are a profit oriented company :).

I suspect a 8c16t will probably be priced like the 6800k at most and price the 6c12t against the 7700k. But i doubt 6c12t will win any singlethreaded benchmark agaist a 770k as it may take a 4c8t zen to be able to clock high enough to undo Intel's 10-15% ipc advantage.
 

Rifter

Lifer
Oct 9, 1999
11,522
751
126
An 8c16t Ryzen will compete with the 8c16t Intel equivalent, not a $350 7700k. Your entire post is invalid.


Your post is invalid because we do not have any pricing info yet. If AMD prices the 8c/16t CPU close to the 7700k then it IS competing against it, period.
 

swilli89

Golden Member
Mar 23, 2010
1,558
1,181
136
I think AMD would market the top-end Ryzen against the i7-7700k. Basically a "look moar cores" argument since they know they can't match both the IPC and clockspeed of a chip like that, nor do they want to shoot for the astronomical prices or platform features of X99 or X299.

They'll bench vs the 6900k as a pissing contest but when it comes to pricing/marketing, the primary target is Z270/7700k.
Well the good part is we will know for sure in about a month what their plan is. But like most businesses they will deploy a marketing segmentation strategy to both increase revenue AND increase their corporate gross margin which is around 32% right now. Lisa stated that longer term they want 40% margin. In order to do that they need to keep operating expenses flat while increasing product profit margin.

They will most definitely target z270 (I Think x370 naming points to this) but they will also target Intel's "HEDT" market. Every public demo has been against the latter platform.

I think we see an 8C/16T unlocked SKU at 3.4/3.8 at $400-450 and we see a top binned 8C/16T 3.6/4.0 Ryzen around $500-600. As its been beat to death already at this point I just want to remind everyone that they don't have to get the entire market! They just need to carve out a nice portion for now. For sure some consumers and businesses will still purchase a 6900K @ $1100. But a sizeable portion of them will definitely opt to save 50% for a basically identically performing top Ryzen chip/platform.
 

beginner99

Diamond Member
Jun 2, 2009
5,208
1,580
136

Lets use every single game Anandtech tested on the 7700K.


So you're flat out wrong, the 7700K doesn't blow ANY 6 or 8 core CPU out of the water. The only thing "HEDT" CPUs accomplish here is being much more expensive WHICH Ryzen should have an answer for in the form of their lowest end $400 (my guess) 8c/16T CPU.

These game test are mostly useless because they are single-player GPU limited benches. So it only applies to people that never play multi-player. However where you will see CPU differences is for example battlefield 64 player maps. This is where these modern games get CPU limited. Or Starcraft 2 4v4 on fast Map. In the later case I would bet quiet a lot of money that the 7700k will win. In case of BF1 or BF4 I don't know but my point stands that these benches are useless as proof for your point. It like saying I'm comparing a ferrari with a ford ka. Both start the race at 30mph and aren't allowed to go faster than 30mph. Yeah, it will be a tie and then claiming they are equivalent is stupid.
 

tamz_msc

Diamond Member
Jan 5, 2017
3,726
3,560
136
If all you want to do is test the CPU's single threaded performance in games, and get a good indication of multiplayer performance, then just benchmark a CSGO map with lots of bots.
 

Magic Hate Ball

Senior member
Feb 2, 2017
290
250
96
Your post is invalid because we do not have any pricing info yet. If AMD prices the 8c/16t CPU close to the 7700k then it IS competing against it, period.

I'm leaning towards the price being way closer to the 7700K than the 6900K because of the language in the comparison Lisa Su was making during the Horizon event.

They kept showing the Ryzen vs the 6700K vs the 6900K and saying "look how poorly the 6700K is keeping up with the Ryzen chip!" then saying "oh see, this ONE THOUSAND DOLLAR CHIP is keeping pace".

They kept reiterating how crazy expensive the Intel chip was for the performance equivalent to the Ryzen while waving the 6700K around as well. This, in my mind, is a signal that the Ryzen 8C/16T will compete more in the 6700K's price territory.

It'll probably land smack dab in the middle, but a man can dream.
 

CatMerc

Golden Member
Jul 16, 2016
1,114
1,149
136
Now your just jamming words together, the infinty fabric is the collection of IP that connects different IP blocks together coherently, it plays no part within a CCX. it also wont have lower latency across the infinity fabric between two CCX's then broadwell-E ringbus. Now better streaming performance could be because of better predictive selection of thread priority but that also doesn't mean its throughput and thus performance is higher.
Not arguing in anyone's favor, but AMD did claim inter-core communication is better thanks to Infinity Fabric. Here is a link: https://youtu.be/vMfNz2SXVLk?t=2m50s

Something about Ryzen, with AMD claiming Infinity fabric playing a part in it, allows it to better juggle the game and a stream at the same time, though throughput is probably similar.
 

DrMrLordX

Lifer
Apr 27, 2000
21,583
10,785
136
If a 8c16t will be having a pissung contest with 6900k, why would they price it against 7700k? They are a profit oriented company :).

That's how they're going to make most of their money: cut into the larger-volume market for high-end mainstream chips (currently the i7-7700k). They have to demo their chip against the 6900k to convince buyers that it will be (at least some of the time) better than the i7-7700k.

That being said I think they will charge more for the top-end chip than Intel charges for the 7700k. The difference here is that I think they're seriously going to try to get Kabylake buyers, not Broadwell-E buyers.

I suspect a 8c16t will probably be priced like the 6800k at most and price the 6c12t against the 7700k. But i doubt 6c12t will win any singlethreaded benchmark agaist a 770k as it may take a 4c8t zen to be able to clock high enough to undo Intel's 10-15% ipc advantage.

6800k is only ~$400, though with platform price, x99 + 6800k will probably cost about the same as X370 + SR7. So I'll agree with you there, partially. The difference here is that I think there aren't enough people actually buying the 6800k for that market segment to really matter.

They will most definitely target z270 (I Think x370 naming points to this) but they will also target Intel's "HEDT" market. Every public demo has been against the latter platform.

See above. They're demoing against HEDT to establish the performance of the product. If all AMD got was the HEDT market buying their chips, sales volume would be way too low (barring the possibility of every 5820k owner out there buying Ryzen; they might be okay in that case).

As its been beat to death already at this point I just want to remind everyone that they don't have to get the entire market!

No, but they do have a great opportunity here, and it would be a shame for them to screw it up by charging too much for the chip.
 
  • Like
Reactions: ZGR

beginner99

Diamond Member
Jun 2, 2009
5,208
1,580
136
No, but they do have a great opportunity here, and it would be a shame for them to screw it up by charging too much for the chip.

Does it really matter? It's a server CPU and if it is a success (performance/watt) it will move tons of units in the server space, so they can price it $600 or higher for consumers and it would not really matter as server sector will buy enough of them anyway.
 

DrMrLordX

Lifer
Apr 27, 2000
21,583
10,785
136
Does it really matter? It's a server CPU and if it is a success (performance/watt) it will move tons of units in the server space, so they can price it $600 or higher for consumers and it would not really matter as server sector will buy enough of them anyway.

A few points:

1). Apparently AMD is already taking shots at the workstation market with the Pro line of Ryzen CPUs (if the leaks are to believed). That line may represent a fair number of potential sales since they could move a large number of units with those SKUs alone. I am not sure of the TCO for professional workstations vs server TCO (haven't dealt with large-scale procurement of professional workstations for a department, etc) so it's hard to say if the pricing of individual CPUs would make that much of a difference. In the server world, it doesn't seem to matter as much what with the software licensing fees and power consumption being bigger factors.

2). Pro CPUs aside, it'll be a bit like the old days in which server guys are swayed by what worked for them in the desktop realm. A bit. Ryzen mass-proliferating throughout enthusiast communities will have quite an impact on professional awareness of the technology platform. Ryzen selling a few thousand units as it hits the floor with a thud will probably disappoint more than a few people in procurement departments.

In the end it'll have more to do with the pricing and performance of the workstation/server parts themselves, it just helps that if they guys/gals responsible for pulling the trigger on large quantities of pro equipment have had the chance to see/use desktop derivatives beforehand. More than a few Opterons back in the k8 days went into server rooms just on the basis that the people involved already knew the qualities of k8.
 

NostaSeronx

Diamond Member
Sep 18, 2011
3,683
1,218
136
Excavator Mk3... the end of Zen(Aspen/K12) coming soon October 2nd, 2017. (or, September 2017)

Bristol Ridge AM4 has outsold Ryzen(summit ridge). So, big things are coming. If it didn't make sense... mk3 is primarily in a CPU SKU.

Edit: Invest in Hype.
 
Last edited:
  • Like
Reactions: itsmydamnation

itsmydamnation

Platinum Member
Feb 6, 2011
2,744
3,080
136
Excavator Mk3... the end of Zen(Aspen/K12) coming soon October 2nd, 2017. (or, September 2017)

Bristol Ridge AM4 has outsold Ryzen(summit ridge). So, big things are coming. If it didn't make sense... mk3 is primarily in a CPU SKU.

Edit: Invest in Hype.
Atleast you make me laugh...........
 

Hitman928

Diamond Member
Apr 15, 2012
5,182
7,633
136
Excavator Mk3... the end of Zen(Aspen/K12) coming soon October 2nd, 2017. (or, September 2017)

Bristol Ridge AM4 has outsold Ryzen(summit ridge). So, big things are coming. If it didn't make sense... mk3 is primarily in a CPU SKU.

Edit: Invest in Hype.

b638d12bbeb913fad87b1642390d3aa6a4548b8999f540b904a5c5484bcd6f75.jpg
 

krumme

Diamond Member
Oct 9, 2009
5,952
1,585
136
This guy has not been wrong yet, and he is saying that Ryzen will cost approximately $500 MAXIMUM. Skip to 4 minute mark.

https://www.youtube.com/watch?v=lyk3lXe5vWA&t=107s
He is using Fottemberg as source. Fottemberg got some oem source. The oem probably dont know the pricing either for sure but the argument that it does not make sense to price it higher looking at the board cost is certainly from the oem. Its a mass market product and its scaled on boards to be so.

We see tons of boards. Anyone remember how it was at K7 launch?
 
  • Like
Reactions: Shamrock

DrMrLordX

Lifer
Apr 27, 2000
21,583
10,785
136
Heh! K7 was pretty sweet, but none of the chipsets supporting it were all that good. Poor memory controllers, IDE data corruption bugs, the works.

At least the boards were fairly inexpensive.
 

itsmydamnation

Platinum Member
Feb 6, 2011
2,744
3,080
136
Heh! K7 was pretty sweet, but none of the chipsets supporting it were all that good. Poor memory controllers, IDE data corruption bugs, the works.

At least the boards were fairly inexpensive.
Had Zero problems with the Asus A7N8X deluxe Rev 2.0 ( yes i remembered that off the top of my head :p )


edit: oh crap was that a dash -E.... i dont remember...lol

but i remember 22x 100 barton 2500XP
ati 9500pro
512 mb ddr400

ahh the good old days :)
 
Last edited:
  • Like
Reactions: Cartel