tamusc said:

---

theterk said:

---

My only qualm with the Tom's article is that everything was tested w/ the Nvidia 1080. If you noticed, as they ramped up resolution, the gap starts to decrease. And the average numbers they listed at 1080p are well above 60fps. I am eagerly awaiting the GPU walkthrough from those guys.

---

The narrowing margins at higher resolutions means that the video card is becoming more and more of the bottleneck rather than the CPU, right?

---

That's how I'd interpret it.

Nice! Look forward to your thoughts on it as it seems to be one of the better boards at launch based on the reviews I've read. Apparently the Asus Crosshair was buggy for a decent number of reviewers, which is disappointing from Asus.

There's no guarantee games will get patched to improve performance, but there definitely seemed to be some consistent issues with memory speeds and other items (like boost being broken on many mobos) that you would think could be fixed via BIOS updates.

http://www.pcworld.com/article/3176053/components-processors/amds-su-says-patches-will-boost-ryzen-gaming-performance-it-will-only-get-better.html

Looks like AMD's other response to the backlash over poor gaming performance is to release the 5 series earlier than planned. Should help the price/performance competition against the quad core Intel chips.

https://www.forbes.com/sites/antonyleather/2017/03/03/amd-announces-ryzen-5-1600x-and-1500x-faster-than-expected-these-are-the-real-intel-beaters/#3730db7255ea

Currently running through my game benches... Metro looked on par w my xeon 1230 v5. Let's see how ashes of singularity does.

Update* currently running 3.8GHz with no additional voltage or loadline adjustment.

The explanation for some poor gaming performance, particularly with some newer games, being caused by issues with SMT due to the scheduler makes sense. Impressive that they saw 17% increased performance when SMT was disabled even though that knocked out have the virtual cores. Same with Windows not properly managing the new cache design on the CPU

I'm back to wait and see mode after spending some time messing with a friend's 7700k system. For what I do, I think I want the better overall system performance and if they can get the gaming performance to within a few percentage points (up from 20% in some instances), I'd be more than fine with that.

Fixing BIOS on motherboards to RAM can run faster than 3000Mhz should help performance as well

Looks like most of what we're seeing is indeed an immature platform and software not taking advantage of new features, neither of which are unexpected.

Looking forward to what theterk has for us.

I'm looking to upgrade from my I5-2500K, which is a home office computer as well as gaming machine.

Do I jump in on the AMD wave or take advantage of the Microcenter discounts on the kaby lake chips and stay with a known product and not fight the growing pains?

I'm really looking forward to getting a Ryzen 7 computer.

CT'97 said:

---

I'm looking to upgrade from my I5-2500K, which is a home office computer as well as gaming machine.

Do I jump in on the AMD wave or take advantage of the Microcenter discounts on the kaby lake chips and stay with a known product and not fight the growing pains?

---

I got a 7700k for $309 off Ebay from Newegg. Could not pass that up. I've had problems with AMD in the past so it was an easy choice for me. AMD always seems to have some sort of weird issues with their drivers. Never had problems with Intel products and this is going back 25+ years. I've had AMD twice and had issues both times (crappy drivers and support was the bulk of it).

Terk bringing the heat like always. Good stuff, man!

tamusc said:

---

So you missed the Pentium that couldn't do math correctly and had to be recalled? Cost Intel half billion fixing that.

I built multiple AMD rigs (K6-2, Athlon, Phenom) for myself and never really had issues with any of them. They were competitive from a performance standpoint while beating Intel on price. That said Intel had a hit with the original Core lineup and then AMD failed miserably with the entire Bullodzer lineup. Since then, we've seen nothing but minor, incremental performance improvements from Intel since they weren't facing any serious competition.

We as consumers need Ryzen to succeed, if nothing else to push Intel. So far, we've seen Intel already respond with big price cuts and push forward yet another small, incremental update to this Fall from next Spring. Two big wins for consumers.

---

Lol...Good point...I'll take my chances on a bug that might happen once every 27000 years.

https://en.wikipedia.org/wiki/Pentium_FDIV_bug

Quote:

---

The severity of the FDIV bug is debated. Intel, producer of the affected chip, claims that the common user would experience it once every 27,000 years while IBM, manufacturer of a chip competing with Intel's Pentium, claims that the common user would experience it once every 24 days. Though rarely encountered by most users (Byte magazine estimated that 1 in 9 billion floating point divides with random parameters would produce inaccurate results),[url=https://en.wikipedia.org/wiki/Pentium_FDIV_bug#cite_note-3][3][/url] both the flaw and Intel's initial handling of the matter were heavily criticized by the tech community. The man who found the bug points out that since it went unnoticed for a year in a popular product, that likely indicates that the bug was less harmful than IBM suggested.

---

It was a 1 in 9 billion bug. There is a reason nobody noticed it.

I put my first AMD chip in back in 94 or 95 and it wasn't terrible. I tried AMD again in the early 2000's and it sucked. Nothing but problems from crappy drivers and crappy support of said drivers. I would rather skip the beta testing and get hardware that works.

my motherboard article got frozen for being too provocative. My analysis comparing Z270's i7-7700k vs X370 1700X was not definitive enough...

oh well.. re-writing it to be a plain jane motherboard review :-/

theterk said:

---

my motherboard article got frozen for being too provocative. My analysis comparing Z270's i7-7700k vs X370 1700X was not definitive enough...

oh well.. re-writing it to be a plain jane motherboard review :-/

---

Too provacative? Seriously?? Maybe you can slip us the first draft after your revised one is published

Either way, looking forward to reading it!

I'll give you driver issues for ATI/AMD graphics cards for sure, but I never ran into driver issues with the chipsets or CPUs that were any worse than ones I ran into with Intel systems at the time.

And do you think the Pentium issue really wasn't a problem? Intel apparently thought it was a problem since they did a recall and spend almost half a billion cleaning up the mess. Companies don't do that unless there's actually a problem.

While waiting on Terk's updated review to be posted, found one that has a different take that brings up some interesting points.

http://www.pcworld.com/article/3179208/gaming/4k-gaming-tested-amd-ryzen-and-geforce-gtx-1080-ti-for-less-than-intels-cheapest-8-core-chip.html

grr, that just ticks me off.... similar things that I wrote.

since they didn't buy this article (but they bought the watered down one), here's my Ashes of the Singulartiy data

*edit* *edit* here it is:

Ugh, that stinks. I wondered if that was kind of where you were going with your original article.

it was, as you can see in the chart, using the GTX 970 I see very comparable performance out of both processors at "High". when you get to Crazy, my system squeaks by the i7-7700k system using the GTX 970.

The Xeon data is carry over from my C236 articles and I think that was a Skylake processor, not Kaby Lake

Yeah, I think that chart shows that it might be short sighted to judge Ryzen based purely on framerates at low graphical settings.

While it's generally true that at lower resolutions, the CPU tends to be the bottleneck, the system isn't entirely limited at higher resolutions by the GPU as some think. There's a reason why people running 4K rigs use the Intel hexa and octacore CPUs. The Intel quadcores can certainly ramp up the speeds to get better single thread performance, but at some point when jumping up to 4k resolutions those same quad core CPUs just can't keep up.

In those instances, I think Ryzen will be more competitive as the platform matures.

Here's a good look at the issue with Windows 10 Scheduler that's causing performance issues with SMT and CPU cache on Ryzen. Another part of the growing pains of a new platform, but should provide a decent boost to Ryzen when it's fixed.

As they point out in the article, Intel had similar issues when they first introduced their own version of SMT.

http://wccftech.com/amd-ryzen-performance-negatively-affected-windows-10-scheduler-bug/

Also props to AMD for not going cheap with the CPU package construction like Intel did a while back.

http://wccftech.com/amd-ryzen-delided-gold-plated-solder-silicone-protected-caps-confirmed-cheap-tim/?utm_source=wccftech&utm_medium=related

Interesting look at the Windows 10 Scheduler issue with Ryzen.

https://www.pcper.com/reviews/Processors/AMD-Ryzen-and-Windows-10-Scheduler-No-Silver-Bullet

TLDR article summary - It looks like that might have disproven that there is indeed a problem with the Windows 10 scheduler based on their testing.

There have been demonstrated real world performance gains in games when SMT was disabled, which is why people thought that it was Windows 10 Scheduler shunting gaming threads to the logical, rather than primary cores, which have lower performance. Those performance gains with SMT disabled combined with this article make me think again that it's more an issue with an immature platform that hopefully can be addressed with some BIOS updates and software patches, rather than something endemic to the CPU architecture itself.

I really think the Ryzen 5 series could be interesting, particular the two hexacore chips, at those price points.

https://arstechnica.com/gadgets/2017/03/amds-zen-goes-mainstream-with-ryzen-5-4-cores-8-threads-from-169/

Also saw that our very own theterk's review was finally posted!

http://www.tomshardware.com/reviews/asrock-x370-taichi-amd-ryzen-am4-atx-motherboard,4967.html

Can't wait to read it when I get some time later today!

yep, had to sweet talk the Senior Editor to push it forward instead of a Saturday release (who reads those?).

Currently defending my honor in the comments section. It is clear that no one likes my GTX 970 data. Maybe they'll like SLId 970s? My next motherboard review is due out in 2 weeks

Don't let the trolls get you down, I thought it was a good review. What's your next review on?

you choose! I was going to do Gigabyte next to not show bias. Also en route are Biostar and MSI samples.

Personally I'd like your thoughts in that Gigabyte Aorus, but again baffled by the choice to put only one m.2 drive slot on it.

Personally I was really looking at the Asus Crosshair Hero, but again, only one m.2 slot when the comparable Intel version has two! Admittedly, I've been a long time Asus fan, but I've also seen mixed reviews on the Crosshair. The Taichi you reviewed is the probably the one I'm looking at now, with the Aorus next.

This could be an interesting move for AMD...

Rumors of a 16 core/32 thread desktop CPU

https://wccftech.com/amd-working-16-core-ryzen-cpu/

tamusc said:

---

This could be an interesting move for AMD...

Rumors of a 16 core/32 thread desktop CPU

https://wccftech.com/amd-working-16-core-ryzen-cpu/

---

A little more info on this one. Looks to be a new platform that will add quad channel RAM in addition to more cores and maybe a few other features. Think I'll see if my 3570k can limp along for a few more months...

http://www.tweaktown.com/news/56757/amd-ryzen-cpu-16c-32t-coming-teased-999/index.html

Pretty good read. Seems to line up with the array of articles and reviews I've read.

http://www.pcgamer.com/amd-ryzen-state-of-the-union-address/?utm_content=buffere798f&utm_medium=social&utm_source=facebook&utm_campaign=buffer_pcgamerfb