Discussion AMD Ryzen MegaThread! FAQ and Resources

[-Fran-]

I dont even understand why you would get the 1600x over 1600 since the x costs more in of itself and additionally the 1600 comes with the cooler which is enough for overclocks to about 3.8-4 Ghz thats for any of the x versions. Only reason would be extreme overclocks.

I have my suspicions that AMD might have been binning the 1600X with working 1700X or 1800X'es. The power figures for them look in-line with the power characteristics the 1800X showed, even more than the 1700. That fully enabled L3 is also a good hint on the binning.

But, I have no proof and only a hunch. That is why I'd gamble with a 1600X instead of the 1600, but you're correct at "face value". The 1600 is the better deal over the 1600X. The difference seems to be mostly around the wraith and the extra 100Mhz out of the box (for the clock-turbo thingy).

Cheers!

 

[salgado18]

Spoiler

I dont even understand why you would get the 1600x over 1600 since the x costs more in of itself and additionally the 1600 comes with the cooler which is enough for overclocks to about 3.8-4 Ghz thats for any of the x versions. Only reason would be extreme overclocks.

I have my suspicions that AMD might have been binning the 1600X with working 1700X or 1800X'es. The power figures for them look in-line with the power characteristics the 1800X showed, even more than the 1700. That fully enabled L3 is also a good hint on the binning.

But, I have no proof and only a hunch. That is why I'd gamble with a 1600X instead of the 1600, but you're correct at "face value". The 1600 is the better deal over the 1600X. The difference seems to be mostly around the wraith and the extra 100Mhz out of the box (for the clock-turbo thingy).

Cheers!

On the other hand, the 1700 has only the stock clocks of a disadvantage over the 1600X, and anyone who's willing to overclock, can get more value out of the 8 core chip. For me, I think that the 1700 is the sweet spot for overclockers, and the 1600X for human beings.

 

[Gon Freecss]

I think the X models will be better for overclocking once AMD sorts the whole overclocking problem. IIRC, they have better power delivery and are pretty much binned for better overclocking.

 

[TechyInAZ]

I dont even understand why you would get the 1600x over 1600 since the x costs more in of itself and additionally the 1600 comes with the cooler which is enough for overclocks to about 3.8-4 Ghz thats for any of the x versions. Only reason would be extreme overclocks.

Well, the 1600 should overclock as well as the 1600X, like the 1700X vs 1700.

Either way I'd pay the extra for the XFR. I personally wouldn't plan to overclock the 1600X much so the XFR would help a lot. (if i was going to buy one that is.)

 

[juanrga]

Too bad Toms didn't have the 1500X numbers in the mix, but other review sites do have them.

There seems to be a *real* advantage to having 6C over 4C, since in everything where the 1500X is close to the i5, the 1600X makes it onto par or better in some games that actually go wide (Crysis 3 being really surprising) and for productivity stuff, the 6C is a steal; hands down.

12 threads vs 4 threads. The 1600X is going to win on well multithreaded code and the i5 on latency-sensitive code

http://www.anandtech.com/show/11244/the-amd-ryzen-5-1600x-vs-core-i5-review-twelve-threads-vs-four/17

http://www.tomshardware.co.uk/amd-ryzen-5-1600x-cpu-review,review-33858-11.html

Intel’s Kaby Lake-based processors beat Ryzen 5 1600X in lightly-threaded applications where they can leverage superior IPC throughput. But the 1600X’s extra cores/threads turn the tables in software well-optimized for multi-core CPUs.

 

[juanrga]

Well the lower end R3 are 4c 8t no?\
Edit:https://www.reddit.com/r/Amd/comments/64e3oe/ryzen_5_1600_cinebench/
Another Cbench leak. Pinch of salt recommended.

Nope (can do this all day)

The R3 are all 4c/4t. With same clocks as the other Ryzen, also unlocked, good IPC, and much lower price than i5's, i'd call them a huge win.

Something tells me we might not have L3 cache or as much as the 1400 does. I even wonder if the R3 is nothing more then the APU but with no iGPU.

All RyZen CPUs use the same 8-core die including the R3 models whose specs were leaked by CanardPC (no-info on L3 amount however)

https://www.overclock3d.net/news/cpu_mainboard/amd_ryzen_1400_and_1200_cpu_specifications_leak/1

The Zen based APUs for desktop have been delayed to 2018

 

[Nope 1151]

Hey Juan. Did you see the toms experiment on memory frequency? 2333 seems like a sweet spot but why would the increase of ram not scale linearly? (That would mean only 1166 if the 1/2 of memory rumor is true)

Is this a limitation by AMD's design by any chance?

 

[juanrga]

Hey Juan. Did you see the toms experiment on memory frequency? 2333 seems like a sweet spot but why would the increase of ram not scale linearly? (That would mean only 1166 if the 1/2 of memory rumor is true)

Is this a limitation by AMD's design by any chance?

Link?

Do you mean why performance does not scale linearly with mem. frequency? It is not a limitation of AMD design, but the expected behavior because IPC is not a linear function of memory frequency. In a first-order model you have an expression like

IPC = 1 / [ SS + BRM + ICM + DCM ]

The terms on the right account for steady state operation of the pipeline, plus the effect of branch misprediction and the effects of instruction cache and data cache misprediction, respectively. Overclocking RAM reduces the terms ICM and DCM, but the rest of terms remain. If you plot a function as this of above considering that (ICM + DCM) is the variable X you find a nonlinear graph of IPC vs X.

There is a point behind which increasing clocks don't produce any performance gain because certain components of the CPU are bottlenecked.

 

[Nope 1151]

http://www.tomshardware.com/reviews/amd-ryzen-5-1600x-cpu-review,5014-2.html
Search: "Ryzen 5 1600X Memory Data Rate "

 

[8350rocks]

I again point out: When CPU performance is not an issue, multi-threading is going to be some percent slower due to the additional overhead involved.

Everyone is forgetting the primary advantage of DX12 is alleviating CPU bottlenecks on CPUs that can not handle a single large rendering thread. Assuming a perfectly optimized DX11 code path, I'd expect DX12 to see performance gain on lower tier CPUs [Pentiums, i3's, and FX CPUs come to mind], and a slight performance loss on more powerful CPUs [most i5's/i7's]. Which coincidentally, is EXACTLY what we see with NVIDIA.

DX11 is higher overhead...and Nvidia's drivers are designed around single threaded games. Their drivers do not benefit from native multithreading like DX12 offers because their software scheduler increases overhead of the API.

It is not that DX12 should be less capable on stronger CPUs...it should be more capable by any account with lower overhead. It is that Nvidia creates more overhead the more threads you can spread your workload across because their driver recombines that work into a main render thread to send the command buffer's work to the GPU.

 

[-Fran-]

Too bad Toms didn't have the 1500X numbers in the mix, but other review sites do have them.

There seems to be a *real* advantage to having 6C over 4C, since in everything where the 1500X is close to the i5, the 1600X makes it onto par or better in some games that actually go wide (Crysis 3 being really surprising) and for productivity stuff, the 6C is a steal; hands down.

12 threads vs 4 threads. The 1600X is going to win on well multithreaded code and the i5 on latency-sensitive code

http://www.anandtech.com/show/11244/the-amd-ryzen-5-1600x-vs-core-i5-review-twelve-threads-vs-four/17

http://www.tomshardware.co.uk/amd-ryzen-5-1600x-cpu-review,review-33858-11.html

Intel’s Kaby Lake-based processors beat Ryzen 5 1600X in lightly-threaded applications where they can leverage superior IPC throughput. But the 1600X’s extra cores/threads turn the tables in software well-optimized for multi-core CPUs.

No, I'm extending that to both 4C/4T and 4C/8T from Intel. That is why I didn't mention "threads" explicitly.

Cheers!

 

[randomizer]

Not many reviews for the 1600 it seems. A bit of a pain. I like having my CPUs come with a HSF. It's one less thing to research.

 

[-Fran-]

This image caught my eye on a second read and is very interesting:

These are the 1:1 for the 1500X and the 7500 respectively:

3.5 / 3.7 GHz - Base/Turbo - 3.4 / 3.8 GHz
16 - PCIe 3.0 Lanes - 16
16 MB - L3 Cache - 6 MB
65 W - TDP - 65 W

At a first glance, that 1 point of difference in CB could mean the 100Mhz of turbo difference. Which in turn means... Ryzen is on par with Kaby Lake?

Cheers!

 

[8350rocks]

This image caught my eye on a second read and is very interesting:

These are the 1:1 for the 1500X and the 7500 respectively:

3.5 / 3.7 GHz - Base/Turbo - 3.4 / 3.8 GHz
16 - PCIe 3.0 Lanes - 16
16 MB - L3 Cache - 6 MB
65 W - TDP - 65 W

At a first glance, that 1 point of difference in CB could mean the 100Mhz of turbo difference. Which in turn means... Ryzen is on par with Kaby Lake?

Cheers!

Nice catch! I skimmed right over ST CB simply because everyone has beaten it to death on the R7s.

 

[salgado18]

This image caught my eye on a second read and is very interesting:

These are the 1:1 for the 1500X and the 7500 respectively:

3.5 / 3.7 GHz - Base/Turbo - 3.4 / 3.8 GHz
16 - PCIe 3.0 Lanes - 16
16 MB - L3 Cache - 6 MB
65 W - TDP - 65 W

At a first glance, that 1 point of difference in CB could mean the 100Mhz of turbo difference. Which in turn means... Ryzen is on par with Kaby Lake?

Cheers!

i5 7500: 155pts / 3.8GHz = 40.8 pts/GHz
R5 1500X: 154pts / 3.7GHz = 41.6 pts/GHz

But:
i5 7400: 141pts / 3.5GHz = 40.3 pts/GHz
i5 7600: 166pts / 4.1GHz = 40.5 pts/GHz

R7 1700: 145pts / 3.7GHz = 39.2 pts/GHz
R7 1800X: 160pts / 4.0GHz = 40 pts/GHz (only turbo)
R5 1800X: 160pts / 4.1GHz = 39 pts/GHz (XFR)
R5 1500X: 154pts / 3.8GHz = 40.5 pts/GHz (XFR)

While i5's are very consistent, Ryzen is not. Also, did XFR got triggered? What was the final clock? Does the 1500X have an XFR range of 100 MHz or 200 MHz? Too many questions for me :/

 

[-Fran-]

The easy way to explain that is temperatures. The 1500X comes with the same Wraith the 1600 and 1700 come with, but they have to use custom cooling for everything else. All that variance could be attributed to Temp and power delivery variations. Look at where the 1700 is in the ST bench: it's the lower ST scorer of the Ryzen bunch.

Also, this is mostly academic, but at least is consistent across the other CB tests in AT's review.

I am just interested in see how Juan and others see those figures. If you ask me, if the ST performance is the same and you see the 7500 being faster in most "single-threaded" (let's be lenient with the term) workloads, how can that difference be explained to the common folk out there looking at the FPS differences?

Juan has been tossing around the "high/low latency" term, but no formal definition on what he actually means, so this could be called "bait" to seek clarification.

Cheers!

 

[8350rocks]

Great video from Wendell, double blind between Intel/AMD systems to attempt to verify the "feel" of gaming without knowing what PC was running: https://www.youtube.com/watch?v=ybF7r4rogHc

EDIT: Interestingly, they take a diplomatic approach about there is not much in terms of difference between them, but they note systems 1 and 3 would hitch in certain games pretty frequently. They had a system with 1800X, a 5960X, and 1700X and a 7700K. They tried all 4 without knowing which was which. They typically cite systems 1 and 3 as having hitches running various games...not going to spoil which ones they were because it was quite interesting.

 

[XBloodyR]

Yea saw it aswell pretty interesting anyway there are too many factors that could play into this smoothness people are talking about thats why testing does not really make it clear. You could use screens with more Hz or only systems with amd gpus not to mention see different outcomes at higher levels of fps which for some gamers as me can be very important because smoothness that comes from fps increases when you are above 200-300 fps and go to like 500 fps is just barely noticeable while you are really in pro skill mode with crazy fast aiming which they for sure did not test for.

 

[juanrga]

This image caught my eye on a second read and is very interesting:

These are the 1:1 for the 1500X and the 7500 respectively:

3.5 / 3.7 GHz - Base/Turbo - 3.4 / 3.8 GHz
16 - PCIe 3.0 Lanes - 16
16 MB - L3 Cache - 6 MB
65 W - TDP - 65 W

At a first glance, that 1 point of difference in CB could mean the 100Mhz of turbo difference. Which in turn means... Ryzen is on par with Kaby Lake?

Not even close. You are seeing XFR on action and the clocks aren't those of above. When clocks are locked RyZen is a good 10% behind Kabylake

 

[juanrga]

Not many reviews for the 1600 it seems. A bit of a pain. I like having my CPUs come with a HSF. It's one less thing to research.

https://videocardz.com/68282/amd-ryzen-5-review-roundup

I count five reviews of the 1600, but many aren't in English and cannot be linked here

 

[Nope 1151]

Not many reviews for the 1600 it seems. A bit of a pain. I like having my CPUs come with a HSF. It's one less thing to research.

https://videocardz.com/68282/amd-ryzen-5-review-roundup

I count five reviews of the 1600, but many aren't in English and cannot be linked here

http://www.tomshardware.com/reviews/amd-ryzen-5-1600x-cpu-review,5014-2.html
@Juan search "Ryzen 5 1600X Memory Data Rate"

Also: https://twitter.com/ASUS_ROG/status/851993055926341633
5.9 on 1600 (Reportedly)

 

[8350rocks]

This image caught my eye on a second read and is very interesting:

These are the 1:1 for the 1500X and the 7500 respectively:

3.5 / 3.7 GHz - Base/Turbo - 3.4 / 3.8 GHz
16 - PCIe 3.0 Lanes - 16
16 MB - L3 Cache - 6 MB
65 W - TDP - 65 W

At a first glance, that 1 point of difference in CB could mean the 100Mhz of turbo difference. Which in turn means... Ryzen is on par with Kaby Lake?

Not even close. You are seeing XFR on action and the clocks aren't those of above. When clocks are locked RyZen is a good 10% behind Kabylake

139/153 = 90.84 = 9.16% gap

 

[jwcrellin]

Running faster memory= 4% perf increase. Full power profile=3-4%. Next month's bios O.C. optimizations= 3-4%. I don't think history will see this as an inferior product.

 

[Nope 1151]

That is.. If those expectations are met... We must not get carried away like a lot of people did when they saw that Ryzen 7 vs 6900k demo at 4k.

 

[-Fran-]

Not even close. You are seeing XFR on action and the clocks aren't those of above. When clocks are locked RyZen is a good 10% behind Kabylake

Spoiler

Isn't XFR 100Mhz? That would mean the 1500X is 1 point behind CB at the same clocks. Hardly a 10% diff there. No matter what you say now, the impression I'm getting is Ryzen is closing the gap with these barrage of updates they're getting. That is not even wishful thinking anymore, there is plenty of proof in the numbers already to make that a fact.

Also, I wonder how the uArch scales with speed in Ryzen. I know the performance should grow linearly, but it's interesting to see that at higher speeds, Ryzen seems to be closing the gap. Am I wrong in seeing more than the evident information?

I wish reviewers would do 3Ghz, 3.5Ghz and 4Ghz in-line tests for it. Well, 4Ghz if Ryzen allows, haha.

Cheers!