Official AMD Ryzen Benchmarks, Reviews, Prices, and Discussion

[french toast]

Command rate - is the delay (in clock cycles) between when chip select is asserted (i.e. the RAM is selected) and commands (i.e. ... The amount of time in cycles when the chip select is executed and the commands can be issued. The lower (1T) the faster the performance, but 2T is used to maintain system stability

atm , 3400Mhz 1T was not possible on Intel's CPU nore AMD (I haven't seen it , please notify me ). This is Highest clock that reported on 1T.

Brilliant thanks

 

[PotatoWithEarsOnSide]

Now would be a good time to re-run that Passmark bench...

 

[.vodka]

Now would be a good time to re-run that Passmark bench...

Yeah man.

AMD getting on top of the bad rumors and putting them to sleep. Love it.

Those tests (physics and prime numbers) would get absolutely crushed at those speeds, there's lots of bandwidth PLUS decent latency (DDR4 3400 C18 is (18/1700)*1000=10.5 ns) and 1T command rate.

 

[itsmydamnation]

Intel definitely knows exactly how Ryzen performs...

Just remember when Nvidia launched the gtx 980ti 1 month before the fury x release...Amd was showing to us fury x beating a titan x...yet Nvidia knew exactly how fury x was going to perform so they released a cut down gtx 980ti just to beat the fury x and dictating the price...

Intel doesnt need to lower prices right now since Ryzen is not out yet...obviously when it launches and if it is fast then they will have to lower prices...

Remember how with the 280 NV knew exactly how RV770 was going to perform.......o wait................
If AMD is still rolling microcode right now that improves memory interop and allows for higher speeds/low latencies then all the ~2400mhz CAS16 tests we see floating around then they dont know exactly how Zen will perform.

 

[LTC8K6]

sandbagged

That's what I think. With all those sensors and that 25mhz granularity, XFR has got to do better than 100mhz.

If it only gets you 100mhz with water, what does it get with air? 50?

 

[french toast]

That's what I think. With all those sensors and that 25mhz granularity, XFR has got to do better than 100mhz.

If it only gets you 100mhz with water, what does it get with air? 50?

You answered that with the first part of your post

 

[LTC8K6]

You answered that with the first part of your post

Yes, but in the back of my mind I think that is indeed all XFR does, which would have been a lot of engineering for not much effect...

 

[GroundZero7]

3400Mhz 1T was not possible on Intel's CPU nore AMD (I haven't seen it , please notify me ). This is Highest clock that reported on 1T.

But..um...memory and latency problems...

 

[swilli89]

Yes, but in the back of my mind I think that is indeed all XFR does, which would have been a lot of engineering for not much effect...

I'm wondering the same thing. 100 Mhz seems like a waste of effort.

However there was probably close to zero actual engineering for this and much more effort went into naming it and marketing it. Cores turbo down and up to all different granular P-states as you know, AMD simply is allowing the processor to check its Temp and power draw and go a notch or 2 higher depending on the situation.

 

[PhonakV30]

I found review that they managed to get DDR4-3733 – 17-19-19-38 1T on Core i7-6700K.

If you look at my Post , therefore theoretical maximum bandwidth
( (2133Mhz*2)*64 / 8 ) / 1024 = 33.328125 Gbyte/s
( (2400Mhz*2)*64 / 8 ) / 1024 = 37.5 Gbyte/s
( (3400Mhz*2)*64 / 8 ) / 1024 = 53.125 Gbyte/s

will AMD Ryzen get at least 50Gb/s with DDR4 3400 18-17-17-37 1T? We will see.

 

[CatMerc]

Yes, but in the back of my mind I think that is indeed all XFR does, which would have been a lot of engineering for not much effect...

You mean next to no engineering? XFR is just a natural extension of the already existing technology found in Bristol Ridge. All they did was allow it to go above the turbo clock according to the stability it detects.

I found review that they managed to get DDR4-3733 – 17-19-19-38 1T on Core i7-6700K.

If you look at my Post , therefore theoretical maximum bandwidth
( (2133Mhz*2)*64 / 8 ) / 1024 = 33.328125 Gbyte/s
( (2400Mhz*2)*64 / 8 ) / 1024 = 37.5 Gbyte/s
( (3400Mhz*2)*64 / 8 ) / 1024 = 53.125 Gbyte/s

will AMD Ryzen get at least 50Gb/s with DDR4 3400 18-17-17-37 1T? We will see.

Those sticks might be running at 3000MT/s, rather than 2133MT/s. The sticks are CMK16GX4M2B3000C15, which is rated at 3000MT/s, and is what reviewers got with their review samples.

http://ranker.sisoftware.net/show_run.php?q=c2ffcdffd9b8d9e4dceadcead9efc9bb86b690f590ad9dbbc8f5c5

Here I tested my 6600K with memory at 3000MT/s 15-17-17-35 2T, and besides getting a similar score, the software detected 4000MT/s for whatever reason. So I don't think Sandra's detection is very good.

 

[Zstream]

4/8 thread @3.4ghz? Depending on your overclocks......Pah what the hell, yea ryzen would crush it!

Yeah, my Haswell won't overclock at all. Guessing I'll stick with my Haswell box

 

[freeskier93]

4/8 thread @3.4ghz? Depending on your overclocks......Pah what the hell, yea ryzen would crush it!

With the right motherboard and a semi decent cooler the 1231 will run at max turbo (3.8GHz) all day long under full load.

That said, I'll be upgrading my 1231 in a couple months once I see full reviews and the early issues get worked out.

 

[LTC8K6]

Yeah, my Haswell won't overclock at all. Guessing I'll stick with my Haswell box

If you have a Z board, your 1231 will do 3.8 with all cores under load. You can just set the multiplier to 38.

Curses! Beaten again!

 

[LTC8K6]

You mean next to no engineering? XFR is just a natural extension of the already existing technology found in Bristol Ridge. All they did was allow it to go above the turbo clock according to the stability it detects.

Oh, well I know nothing about Bristol Ridge at all.

 

[Phynaz]

AMD Ryzen 7 1800X hits 5.2 GHz, breaks benchmarking record

http://www.pcgamer.com/amd-ryzen-7-...source=facebook&utm_campaign=buffer_pcgamerfb

2449 doesn't even make the top 100, let alone breaking the record. Time for me put pcgamer on the too lazy to fact check list.

 

[CatMerc]

2449 doesn't even make the top 100, let alone breaking the record. Time for me put pcgamer on the too lazy to fact check list.

8 core record.
Previous one was held by 5960X at 6.04GHz, scoring 2445.

 

[moonbogg]

If my 1800x hits 4.6 on water, I will be so excited I may end up swallowing my tongue or something. Whatever they clock to, I am committed now and will be using the chip.

 

[KTE]

No one really knows if XFR is a complete gimmick or if technically it allows maximum clocks per cooling.

If the latter is true AND only 100-200MHz is gained, Ryzen is launching right at the limits of its LPP clocking curve.

The sign of that happening would be high stock current usage (>110A), for which, 4pin 12v and standard VRMs won't suffice.

Phenom had this problem with high stock currents killing mobos. I wrecked at least 4 different ones in 4 months.

Same with SCT vs ACT. Having 10W vs 85W budget for a Core must allow far higher clocks, unless something else major impediments this.

Sent from HTC 10
(Opinions are own)

 

[Phynaz]

8 core record.
Previous one was held by 5960X at 6.04GHz, scoring 2445.

I see that, thanks. The article is still wrong in their claim, the say the previous record was 2410. That's why I said they were too lazy to fact check before they published.

 

[unseenmorbidity]

If my 1800x hits 4.6 on water, I will be so excited I may end up swallowing my tongue or something. Whatever they clock to, I am committed now and will be using the chip.

Ehh, I wouldn't hope for more than 4.2GHz.

 

[Magic Hate Ball]

No one really knows if XFR is a complete gimmick or if technically it allows maximum clocks per cooling.

If the latter is true AND only 100-200MHz is gained, Ryzen is launching right at the limits of its LPP clocking curve.

The sign of that happening would be high stock current usage (>110A), for which, 4pin 12v and standard VRMs won't suffice.

Phenom had this problem with high stock currents killing mobos. I wrecked at least 4 different ones in 4 months.

Same with SCT vs ACT. Having 10W vs 85W budget for a Core must allow far higher clocks, unless something else major impediments this.

Sent from HTC 10
(Opinions are own)

I have a feeling that XFR is just a bonus feature of their power saving features for laptop/server chips.

I think all these power sensors and smart clocking are in there for maximum mobile power savings (low end clocking) to make their chips competitive in the biggest market segment. It will also pay dividends in the server market where power consumption is a MAJOR selling point.

I don't think they would want to redesign Zen for Workstation/Mobile/Server... so we just happen to get that feature in desktop as well, but probably implemented in a way for more oomph rather than fewer watts on average.

 

[bjt2]

No one really knows if XFR is a complete gimmick or if technically it allows maximum clocks per cooling.

If the latter is true AND only 100-200MHz is gained, Ryzen is launching right at the limits of its LPP clocking curve.

The sign of that happening would be high stock current usage (>110A), for which, 4pin 12v and standard VRMs won't suffice.

Phenom had this problem with high stock currents killing mobos. I wrecked at least 4 different ones in 4 months.

Same with SCT vs ACT. Having 10W vs 85W budget for a Core must allow far higher clocks, unless something else major impediments this.

Sent from HTC 10
(Opinions are own)

Remember that the CPU must remain under 95W TDP, at stock, and one core clock is limited by temperature and not power...

 

[Minkoff]

XFR probably constantly fills the gap beyond Precision Boost, but within the available TDP (or predefined temperature) and could be dependent on the type of workload. If AMD presentation is anything to go by, it can scale even to 500Mhz with proper cooling.

Here Precision Boost is fixed at 4Ghz

 

[itsmydamnation]

There is a video floating around ( thats NDA is supposed to expire 2nd mar) that explains it, XRF is only 100mhz boost if the CPU detects it still has head room, but it applies to both all core clock and single core clock.