Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Post
  • Reply
Worf
Sep 12, 2017

If only Seth would love me like I love him!

got it right ok that makes more sense, bc in my head i was comparing it to those 6th gen i5s and i guess i didnt realize it was 4c/8t

dang thats p badass

Adbot
ADBOT LOVES YOU

gradenko_2000
Oct 5, 2010

HELL SERPENT
Lipstick Apathy
AMD has offered 4c/8t in the past with the Ryzen 5 1400*, but even that had a launch SRP of 169 USD and was still on the 14nm Zen 1 architecture, so performance wasn't that great compared to its Intel contemporaries, and the price still wasn't that cheap either

but then you look at the 3600, which is 6c/12t, on the Zen 2 arch, and at 199 USD was still going for cheaper than an i5-6600K, and the value was just all there as long as you could bear the price

or compare to the i3-9100F, with a 122 USD SRP without hyperthreading, and the low end of the market is pretty blown-out in favor of AMD (assuming they can meet demand) that Intel had to make their 10th gen i3's all 4c/8t to try and compete




* it bugs me that it was the Ryzen 3 1200, which is the 4c without SMT, that got the 12nm refresh, and not this one.

Klyith
Aug 3, 2007

GBS Pledge Week

gradenko_2000 posted:

* it bugs me that it was the Ryzen 3 1200, which is the 4c without SMT, that got the 12nm refresh, and not this one.

AMD, shortly after ryzen launched and was an immediate success, publicly said to press (and investors) that they were going to be a lot more cautious about the low end, because in the past they ended up competing with themselves. They don't want to be the bargain system, and part of that is by avoiding making the bargain too good.

Thus dropping the 1400 4c/8t with Zen+ and not bringing it back until now. If intel hadn't put hyperthreading back into their mainstream chips AMD probably wouldn't have released the 4c/8t Zen 2 parts either.


Also note how the 3700X hasn't gotten any price drops like the 3600 or 3900X. It's still $300, because Intel didn't put any pressure on it whatsoever. Old AMD would have been :derp:price cuts on everything!:derp: as soon as Intel announced the 10 series. (Sad for me, because putting a 3700X into my X370 is my plan for keeping this system until we're 2 or 3 years into DDR5 systems.)

Vadoc
Dec 31, 2007

Guess who made waffles...


Maybe there will be a drop when the 3750 and 3850 refreshes come out?

SamDabbers
May 26, 2003



Klyith posted:

(Sad for me, because putting a 3700X into my X370 is my plan for keeping this system until we're 2 or 3 years into DDR5 systems.)

This was my plan too, but I went with a 3900X because it's the best $/core value of the supported options for X370.

Risky Bisquick
Jan 18, 2008

PLEASE LET ME WRITE YOUR VICTIM IMPACT STATEMENT SO I CAN FURTHER DEMONSTRATE THE CALAMITY THAT IS OUR JUSTICE SYSTEM.



Buglord
I'm was looking to score a 3900x for my b350 when the 4000 series drops, or alternatively grab a 12c 4000 series and b550 board for more nvme drives. Seems like everyone is eyeing the 3900x which may keep prices high but lets see what happens.

I'm wondering if they will have multiple skus for 12/16c depending on clock speed.

Gonna Send It
Jul 8, 2010

Generic Monk posted:

I just built a new PC with a 3700X in it and bought 3200mhz RAM becuase I heard somewhere that was the 'sweet spot' for performance. Enabled the XMP profile which seemed to work, but I got very occasional random restarts when the PC's sleeping that clear all the BIOS settings which is irritating.

Try setting it to the XMP settings manually instead of the profile. I had to do this and I've seen lots of reports where others had to do this as well.

mdxi
Mar 13, 2006

to JERK OFF is to be close to GOD... only with SPURTING

Since there's so much talk about 500-series mobos in comparison to B450 models, here's some data that I gathered today on that very topic.

As always, all mearsurements are with all threads loaded, and given some minutes to stabilize/saturate.

This was pretty surprising to me, as someone who has a good grasp on hardware until the level where we stop talking about bits and start talking about watts/joules. If anyone has the electrical engineering type chops to provide an explanation, I'd really like to understand why I'm seeing what I'm seeing.

Klyith
Aug 3, 2007

GBS Pledge Week

mdxi posted:

Since there's so much talk about 500-series mobos in comparison to B450 models, here's some data that I gathered today on that very topic.

As always, all mearsurements are with all threads loaded, and given some minutes to stabilize/saturate.

This was pretty surprising to me, as someone who has a good grasp on hardware until the level where we stop talking about bits and start talking about watts/joules. If anyone has the electrical engineering type chops to provide an explanation, I'd really like to understand why I'm seeing what I'm seeing.

It might be the X570, or Gigabyte, or just a newer and more tuned bios. PB2/PPT stuff uses tables of power/frequency/voltage curves to set responses to load from the CPU & mobo, and I'm not sure how much of that info is baked into the CPU, how much in AMD AGESA firmware, and how much is mobo-vendor tuned numbers. Do you know offhand which versions of BIOS each is running?


But it also might be the VRM quality that makes some difference. Your X570 is a 6 phase VRM, the asrock B450s are 3 phase. Any time you have more phases the voltage is going to be steadier and faster responding, which in a manual OC scenario means you can get slightly better frequency at any particular voltage. I don't know how much Precision Boost 2 is able to detect and respond to that stuff, but if so that would be a big difference between the two boards.

If you wanted to do an experiment, you could poke inside the Asrock BIOS to see if you can increase the switching frequency of the VRMs. That would, in theory, help compensate for fewer phases. (And decrease VRM efficiency, so consider the cooling.) If that on its own allows the CPU to increase frequency or decrease power, it would be compelling evidence.

mdxi
Mar 13, 2006

to JERK OFF is to be close to GOD... only with SPURTING

Klyith posted:

Do you know offhand which versions of BIOS each is running?

I do know for the ASRock boards. They're all on 3.50, which corresponds to AGESA 1.0.0.3ABB, from August of last year. ASRock didn't issue another BIOS update until November, and since I had gotten the machines stable I didn't want to perturb anything. It might be worth updating them to AGESA 1.0.0.4B now, though.

EmpyreanFlux
Mar 1, 2013

The AUDACITY! The IMPUDENCE! The unabated NERVE!
A Zen2 refresh can kind of make sense if a second wave can reduce IF latency and improve clocks for noticeable improvements, not merely a launch of the "same" as what they're supposed to be replacing. I think using "XT" is probably easier for AMD than series increments they've managed to corner themselves in, but I could definitely see either, but it's also confusing asf regarding their GPUs. Maybe bring back the memory of XP?

429$ 3900XP (4.7Ghz SC, 4.4Ghz AC) vs 10900K
329$ 3800XP (4.8Ghz SC, 4.5Ghz AC) vs 10700K
229$ 3600XP (4.6Ghz SC, 4.4Ghz AC) vs 10600K

Might be reasonable for a refresh if the IF or latency could be reduced as well and this could all be done on under 1.3v? Beat on price and get a lot closer to matching performance in SC and noticeably winning AC? Then Zen3 drops late 2020 and Intel has spent the entire time with basically no wins going into 2021 so even when Rocket Lake launches it gets kind of ignored as every just waits for Zen4 (Which AMD could do another XP launch to wrap up on AM4 X570/B550 before AM5 drops).

Wrar
Sep 9, 2002


Soiled Meat

EmpyreanFlux posted:

A Zen2 refresh can kind of make sense if a second wave can reduce IF latency and improve clocks for noticeable improvements, not merely a launch of the "same" as what they're supposed to be replacing. I think using "XT" is probably easier for AMD than series increments they've managed to corner themselves in, but I could definitely see either, but it's also confusing asf regarding their GPUs. Maybe bring back the memory of XP?

429$ 3900XP (4.7Ghz SC, 4.4Ghz AC) vs 10900K
329$ 3800XP (4.8Ghz SC, 4.5Ghz AC) vs 10700K
229$ 3600XP (4.6Ghz SC, 4.4Ghz AC) vs 10600K

Might be reasonable for a refresh if the IF or latency could be reduced as well and this could all be done on under 1.3v? Beat on price and get a lot closer to matching performance in SC and noticeably winning AC? Then Zen3 drops late 2020 and Intel has spent the entire time with basically no wins going into 2021 so even when Rocket Lake launches it gets kind of ignored as every just waits for Zen4 (Which AMD could do another XP launch to wrap up on AM4 X570/B550 before AM5 drops).

XT is a clear reference to Radeons, who have used XT as a faster designation since 9xxx in the early 2000s, like Nvidia does with Ti.

HalloKitty
Sep 30, 2005

Adjust the bass and let the Alpine blast

Wrar posted:

XT is a clear reference to Radeons, who have used XT as a faster designation since 9xxx in the early 2000s, like Nvidia does with Ti.

NVIDIA have also used XT.

XP would be neat simply because, I don't know, Athlon XP. Still, XT is more likely.. or incrementing the model number by 25, which would be a cleaner solution in my opinion

EmpyreanFlux
Mar 1, 2013

The AUDACITY! The IMPUDENCE! The unabated NERVE!

Wrar posted:

XT is a clear reference to Radeons, who have used XT as a faster designation since 9xxx in the early 2000s, like Nvidia does with Ti.

It's why I think XT is a bad idea, it's easy potential for confusion. I mean, we wouldn't be that far from having a 5700XT paired with your 5700 XT.

CaptainSarcastic
Jul 6, 2013



Ryzen 5a 3600Xtreme2 Obsidian Edition

Mr.Radar
Nov 5, 2005

You guys aren't going to believe this, but that guy is our games teacher.

CaptainSarcastic posted:

Ryzen 5a 3600Xtreme2 Obsidian Edition

Forgot the AF.

FuturePastNow
May 19, 2014


gotta put the X in front of the number, X3700X

Fabulousity
Dec 29, 2008

Number One I order you to take a number two.

They should do something more emblematic of the times and use unicode symbols or emojis. AMD Ryzen R5 3600 ᔑ◕‿◕✿ᔐ, AMD Ryzen R9 3900 ☜(゚ヮ゚☜), etc.

Fabulousity fucked around with this message at 05:58 on May 25, 2020

gradenko_2000
Oct 5, 2010

HELL SERPENT
Lipstick Apathy
Ryzen 5 3600/r/AMD

mdxi
Mar 13, 2006

to JERK OFF is to be close to GOD... only with SPURTING

Ryzen 4300
Ryzen 4600🍆
Ryzen 4950🍆💦

Ep🍑c

All the ideas I came up with for Threadripper were just annoying when written out.

TheFluff
Dec 13, 2006

FRIENDS, LISTEN TO ME
I AM A SEAGULL
OF WEALTH AND TASTE

Klyith posted:

1. Get memtest as a faster way to test if your ram is good than waiting for random crashes. If you have unstable memory memtest will generally tell you in under 15 minutes. (Also it's a good idea to test new ram with it even if you aren't having frequent crashes, because if the ram has a defective spot memtest will find it.)

Memtest86 is fine for finding hardware problems with RAM sticks but not so great at finding instabilities caused by overclocking. For that, Google stressapptest (free, runs under WSL or Linux on a USB stick) or Karhu RAMTest ( :10bux: ) are better options.

Testing memory OC stability is a great way to drive yourself completely bananas within a week though. It takes for loving ever and sometimes settings that just passed 12 hours of testing fine will throw an error within the first five minutes the next day.

TheFluff fucked around with this message at 10:46 on May 25, 2020

EmpyreanFlux
Mar 1, 2013

The AUDACITY! The IMPUDENCE! The unabated NERVE!
People on Chiphell are saying
3900XT: 4.8Ghz Boost, 4.1Ghz Base
3800XT: 4.7Ghz Boost, 4.2Ghz Base
3600XT: 4.7Ghz Boost, 4.0Ghz Base

And 1usmus says 2000Mhz FCLK would be average OC potential. I feel like this is Zen1/Zen1+ all over again, where this is basically just porting Zen2 or N7P (or even N6) and getting the clocks up purely because of dark silicon. I mean if Zen2+ can hit these clocks (and is actually real) I don't see why Zen3 won't hit similar if not better? I don't expect any kind of regression.

All predicated on this like, actually being real.

Drakhoran
Oct 21, 2012

I presume the 3000 series refresh means we won't see 4000 series desktop chips until late Q4 at the earliest?

gradenko_2000
Oct 5, 2010

HELL SERPENT
Lipstick Apathy

Drakhoran posted:

I presume the 3000 series refresh means we won't see 4000 series desktop chips until late Q4 at the earliest?

if they're making the official announcement in June, and availability in July (based on WCCF, so grain of salt), still leaves room for Zen 3 CPUs in Q3.

I feel like this XT refresh is a way to compete harder with Comet Lake, squeeze out some more sales from people right before the Zen 3's release, and take advantage of process improvements without just selling the chips under current SKUs (I've been hearing that more recent Zen 2 CPUs are more capable of hitting their advertised boost clocks than on-release ones)

Arzachel
May 12, 2012
A refresh this late would be spicy because it implies that Zen3 will perform favorably compared to a discounted Zen2+ with ~10% higher clocks.

v1ld
Apr 16, 2012

Arzachel posted:

A refresh this late would be spicy because it implies that Zen3 will perform favorably compared to a discounted Zen2+ with ~10% higher clocks.

Yeah, this was my reaction to those higher clocks as well. They have to avoid stealing the thunder from their own next release - so that's quite spicy indeed if true.

ConanTheLibrarian
Aug 13, 2004


dis buch is late
Fallen Rib

gradenko_2000 posted:

if they're making the official announcement in June, and availability in July (based on WCCF, so grain of salt), still leaves room for Zen 3 CPUs in Q3.

I feel like this XT refresh is a way to compete harder with Comet Lake, squeeze out some more sales from people right before the Zen 3's release, and take advantage of process improvements without just selling the chips under current SKUs (I've been hearing that more recent Zen 2 CPUs are more capable of hitting their advertised boost clocks than on-release ones)
July is Q3. I think anyone expecting Zen3 before late this year will be disappointed. I agree that it doesn't seem likely the refresh is anything other than AMD taking advantage of the improvements TSMC made to Zen2's process to do some more binning. Seems like the kind of thing we'll see more and more of as CPUs are released with clocks that push the limit of the currently achievable frequencies.

Otakufag
Aug 23, 2004
I hope they do that ccx crap they did with 3300x to improve gaming perf

EmpyreanFlux
Mar 1, 2013

The AUDACITY! The IMPUDENCE! The unabated NERVE!

Otakufag posted:

I hope they do that ccx crap they did with 3300x to improve gaming perf

The 3300X doesn't really have improved gaming performance just not hindered gaming performance compared to the 3100. It's basically neck and neck with a 3600/3700X in benchmarks, so it's clear the bottleneck is both the fragmented L3 and latency to system DRAM, and that can only really be tackled in 1) a redesign and 2) higher fabric clocks.

HalloKitty
Sep 30, 2005

Adjust the bass and let the Alpine blast
A 3900X with 200MHz more at the top end, 300 more base would be probably enough to make me consider that instead of waiting for a cheap 3950X. drat.

EmpyreanFlux
Mar 1, 2013

The AUDACITY! The IMPUDENCE! The unabated NERVE!

HalloKitty posted:

A 3900X with 200MHz more at the top end, 300 more base would be probably enough to make me consider that instead of waiting for a cheap 3950X. drat.

I'm just imagining the depreciation on Zen2/2+ vs brand new Zen3 and ohboy is that going to be rough. Like the 3600 already goes for as little as 160$ sometimes, legit watching Intel squirm as a 3600 drops to i3 Comet Lake pricing and the 3700X competes with i5s.

CommieGIR
Aug 22, 2006

The blue glow is a feature, not a bug


Pillbug
https://twitter.com/CommieGIR/status/1264915271669764096?s=19

GRINDCORE MEGGIDO
Feb 28, 1985


Is Stallman onboard with AMD yet?

CommieGIR
Aug 22, 2006

The blue glow is a feature, not a bug


Pillbug

GRINDCORE MEGGIDO posted:

Is Stallman onboard with AMD yet?

I think his name is kinda mud after the falling out with the FOSS and MIT over his comments on Epstein and underage girls

Fantastic Foreskin
Jan 6, 2013

A golden helix streaked skyward from the Helvault. A thunderous explosion shattered the silver monolith and Avacyn emerged, free from her prison at last.

CommieGIR posted:

I think his name is kinda mud after the falling out with the FOSS and MIT over his comments on Epstein and underage girls

Woof. Those are quite the comments.

Craptacular!
Jul 9, 2001

Fuck the DH

GRINDCORE MEGGIDO posted:

Is Stallman onboard with AMD yet?

Last I checked he resists the idea of security coprocessor due to lack of documentation. He sticks to quirky laptops made of parts of mostly Chinese origin.

He’s also less relevant than ever, though that’s hardly news as it’s been the case for most of the life of this forum.

Drakhoran
Oct 21, 2012

Greg Kroah-Hartman is also getting a Threadripper, courtesy of Wendell:

https://www.youtube.com/watch?v=37RP9I3_TBo

iv46vi
Apr 2, 2010
It would be nice to have the spec bump CPUs come out concurrently with B550 motherboards hitting the market.

I want to see the glorious performance per watt per dollar decision tables that will be floating around.

GRINDCORE MEGGIDO
Feb 28, 1985


CommieGIR posted:

I think his name is kinda mud after the falling out with the FOSS and MIT over his comments on Epstein and underage girls
Oh poo poo, he was just a weird foot eating gnome to me.

Adbot
ADBOT LOVES YOU

K8.0
Feb 26, 2004

Her Majesty's 56th Regiment of Foot
Stallman's always been the sort of shithead who believes that anyone who believes exactly what he believes at this specific moment is inherently good and like him could never have done anything wrong, and anyone who disagrees is completely evil and could never have done anything good. He's a really reprehensible piece of poo poo, and has an comically authoritarian view of free software.

  • 1
  • 2
  • 3
  • 4
  • 5
  • Post
  • Reply