Ryzen 3000 boost clock controversy - der8auer publishes his survey results, not a good look for AMD

Meeho

[H]ardness Supreme
Joined
Aug 16, 2010
Messages
4,754
I find it interesting how bent out of shape people are getting over what is arguably a very minor issue.... How often do you guys even use single core loads other than benchmarks? Stop staring at the clock speeds and use your computers. Ryzen 3000 is blazing fast even if you are 100mhz short on single thread loads.
I agree. AMD should market the 3950X as a 5 GHz CPU. It looks better, and the performance will be great either way.
 

Derangel

[H]ard as it Gets
Joined
Jan 31, 2008
Messages
18,496
I find it interesting how bent out of shape people are getting over what is arguably a very minor issue.... How often do you guys even use single core loads other than benchmarks? Stop staring at the clock speeds and use your computers. Ryzen 3000 is blazing fast even if you are 100mhz short on single thread loads.
Point out a single person that has called out AMD for this that is even remotely "getting bent out of shape". The only people getting bent out of shape are the people pissing and moaning about others sharing their opinion.
 

elite.mafia

Broke Back [H]
Joined
Aug 23, 2004
Messages
12,130
Point out a single person that has called out AMD for this that is even remotely "getting bent out of shape". The only people getting bent out of shape are the people pissing and moaning about others sharing their opinion.
Dude theres 8 pages of people resorting to personal attacks because of their opinions on this issue...
 

ccityinstaller

[H]ardness Supreme
Joined
Feb 23, 2007
Messages
4,236
Just because this isn't on the same level as Soundblaster driver issues on a 486 doesn't mean it's not a problem. My patience for computer hardware to just work as advertised has been molded over several generations of products now. If, at any point, your customer is wondering if they need to RMA what they have because they cannot figure out if it's a problem with the product or just working as designed, the company has failed in one way or another.
Dude, we get it. Sell the CPU and buy something else. No matter the fix coming, there is always the possibility you just hit the bottom of the bell curve and got a poor sample.

And before your friends come back, I am not defending AMD here. I would say the same thing if you have a 9900k that can't boost to 5Ghz. They do exist you know.
 

NKD

[H]ardness Supreme
Joined
Aug 26, 2007
Messages
7,785
It's not hyperbole and the difference isn't 25mhz, that's just one segment that everybody is latching to to use as a best case scenario counter argument to minimize the issue/bug. For my 3700x its 125mhz. More for some unlucky souls esp with the 3900x.

Also, it seems good is coming from it. Silence from amd for almost exactly 2 months followed by a sudden update 48 hours after this story finally broke the mainstream. I'm glad AMD is reacting.

Some of you have only noticed this for 2 days, those of us with the hardware have been spending hours and hours trying to fix it since early july.
To think that AMD all of sudden has a solution in few days is nonsense. I always said AMD isn't blind to community feedback and people complaining all over reddit. I assure you AMD has been probably looking in to this for a while. They didn't just comeup with a solution and identified a fix in a few days. They pretty much made a comment when they were close to a fix. That is how it always happens. They did that with destiny 2 update and other errors that were happening as well. Where they made a comment when they were close to a fix. No magic happened here. They announced when they were ready that's all there is to it.
 

elite.mafia

Broke Back [H]
Joined
Aug 23, 2004
Messages
12,130
To think that AMD all of sudden has a solution in few days is nonsense. I always said AMD isn't blind to community feedback and people complaining all over reddit. I assure you AMD has been probably looking in to this for a while. They didn't just comeup with a solution and identified a fix in a few days. They pretty much made a comment when they were close to a fix. That is how it always happens. They did that with destiny 2 update and other errors that were happening as well. Where they made a comment when they were close to a fix. No magic happened here. They announced when they were ready that's all there is to it.
Yep. It may be a 100% fix. It may not be.

One thing I wonder about these chips is the way they boost, it's possible that the reporting/monitoring software isn't accurately displaying clock speeds. I know that cpu-z, hwmonitor and ryzen master all report different speeds, so it seems very likely that because of how fast the CPU changes clock speeds, it may not be accurately reading the clocks.
 

ccityinstaller

[H]ardness Supreme
Joined
Feb 23, 2007
Messages
4,236
I find it interesting how bent out of shape people are getting over what is arguably a very minor issue.... How often do you guys even use single core loads other than benchmarks? Stop staring at the clock speeds and use your computers. Stop running benchmarks over and over. Ryzen 3000 is blazing fast even if you are 100mhz short on single thread loads. Personally I'm very satisfied with my 3700x even if it won't hit 4.4 - as if 4375 is really not the same thing. and still really don't see the big deal and why people are obsessing over single threaded peak boost speeds. What workloads even need that extra 100mhz anyways? Everything is multi threaded these days
People gotta have something to complain about..:eek:
 

SilverSliver

Beat It To Deformation
Joined
Feb 23, 2004
Messages
11,230
I find it interesting how bent out of shape people are getting over what is arguably a very minor issue.... How often do you guys even use single core loads other than benchmarks? Stop staring at the clock speeds and use your computers. Stop running benchmarks over and over. Ryzen 3000 is blazing fast even if you are 100mhz short on single thread loads. Personally I'm very satisfied with my 3700x even if it won't hit 4.4 - as if 4375 is really not the same thing. and still really don't see the big deal and why people are obsessing over single threaded peak boost speeds. What workloads even need that extra 100mhz anyways? Everything is multi threaded these days
100mhz? 3900x - multicore loads are over 500mhz less than max boost. Lightly threaded gaming loads are over 300mhz less than max boost.

At what threshold do you believe it is acceptable to complain? 1ghz?
 

NKD

[H]ardness Supreme
Joined
Aug 26, 2007
Messages
7,785
100mhz? 3900x - multicore loads are over 500mhz less than max boost. Lightly threaded gaming loads are over 300mhz less than max boost.

At what threshold do you believe it is acceptable to complain? 1ghz?
Congrats you have failed in the first place. For the millionth time max boost is frickin single core boost not multicore. I can not believe how misinformed some here are. AMD has never stated in any of their ryzen processors that max boost is all core. It has always been single core.
 

elite.mafia

Broke Back [H]
Joined
Aug 23, 2004
Messages
12,130
100mhz? 3900x - multicore loads are over 500mhz less than max boost. Lightly threaded gaming loads are over 300mhz less than max boost.

At what threshold do you believe it is acceptable to complain? 1ghz?
I said previously and in my post, the ONLY CPU that has any "major" issue with not hitting advertised speeds is the 3900x. And in derbauer's video, the majority of the CPU's were hitting between 4.45 and 4.6... There are some outliers where they are seeing very far away from the advertised max boost clock, but there could be a number of reasons for that and it's very hard to determine exactly what is going on from anonymous survey results.

and why the hell do you expect gaming loads to be at 4.6 ghz on a 3900x? Are you serious?
 

Dan_D

[H]ard as it Gets
Joined
Feb 9, 2002
Messages
55,504
100mhz? 3900x - multicore loads are over 500mhz less than max boost. Lightly threaded gaming loads are over 300mhz less than max boost.

At what threshold do you believe it is acceptable to complain? 1ghz?
From the beginning, AMD has been very straight forward about the fact that heavily multi-threaded workloads would be substantially less than the maximum boost clocks. In the case of the 3900X, it is about 500MHz. AMD's reviewer's guide even points out that what you'll see is roughly a 4.15GHz core clock on all the cores using the 3900X as an example. Its worth pointing out that Intel CPU's show reduced clocks in multi-threaded applications compared to single-threaded tests. Obviously, the discrepancy is much larger on the AMD side, but as long as AMD has been honest about the difference (which they have) I don't see the issue. It's also worth noting that AMD is still considerably faster than Intel in productivity applications. I think people complaining about single-core clocking related to maximum boost clocks not being as advertised, makes some sense. Simply put, there are people who aren't seeing the maximum advertised boost clocks under conditions which should allow for it.

I understand the people who are upset over their single-core / single-threaded boost clocks. What I don't understand are the people who complain about the all core values when what we see, even out of the "bad" examples of Ryzen 3000 series CPU's match up with the information provided by AMD and reviewers.

Congrats you have failed in the first place. For the millionth time max boost is frickin single core boost not multicore. I can not believe how misinformed some here are. AMD has never stated in any of their ryzen processors that max boost is all core. It has always been single core.
Exactly.

I said previously and in my post, the ONLY CPU that has any "major" issue with not hitting advertised speeds is the 3900x. And in derbauer's video, the majority of the CPU's were hitting between 4.45 and 4.6... There are some outliers where they are seeing very far away from the advertised max boost clock, but there could be a number of reasons for that and it's very hard to determine exactly what is going on from anonymous survey results.

and why the hell do you expect gaming loads to be at 4.6 ghz on a 3900x? Are you serious?
I've seen Ryzen 7 3700X's not reach their advertised boost clocks. They are usually only off by about 80MHz or so, but its a problem for a lot of Ryzen CPU's.
 

Mode13

Gawd
Joined
Jun 11, 2018
Messages
746
I've had a 3900X since 2 weeks after launch and whatever bios you decided to run had a huge effect on it. Some had to upgrade to a newer bios to fix a issue that was happening and for some that lowered their single core boost speeds by a little bit all core was rarely affected. It's been obvious for quite sometime it would need a bios tweak to fix it for owners and a few boards have the bios setup right and the chip reaches max boost and sometimes even beyond. ASUS seems to have the biggest issue and is taking the longest to issue bios fixes and a good chunk of people with issues are on ASUS boards. But even other manufactures have boards that work properly and some that dont so it's a bit of a mess. The chips are pretty new and you have to expect minor issues that should get corrected over time, they also have 3 generations of chipsets to troubleshoot as well. You guys seem to have no clue how bad it could be, should have dealt with a VIA chipset and a AMD Athlon processor. Then you would know what a issue looks like and how to have patience for a fix.

I did deal with VIA chipsets and still do, I'm an active member on retro forums and enjoy repairing old socket 7 era boards right up to the socket/slot A crap. :p

Anyway, I guess it's hard to convey feelings through the internet. I'm not on the edge of my chair banging my desk angrily for this fix. I think the vast majority of us are not. It's just conversation and enthusiasm to see some higher clocks. Pitchfork is still in the shed! I've been saying from the start it's a microcode update that's needed along with everyone else. I'm just glad AMD has finally said something/anything about the issue.

To think that AMD all of sudden has a solution in few days is nonsense. I always said AMD isn't blind to community feedback and people complaining all over reddit. I assure you AMD has been probably looking in to this for a while. They didn't just comeup with a solution and identified a fix in a few days. They pretty much made a comment when they were close to a fix. That is how it always happens. They did that with destiny 2 update and other errors that were happening as well. Where they made a comment when they were close to a fix. No magic happened here. They announced when they were ready that's all there is to it.
Obviously, no way they can just bang out a whole new AGESA in 48 hours after seeing a video. However, they're surely aware of the derbauer video and ensuing drama, so whether it's coincidence or not, they finally acknowledged the issue by saying something, which is soothing... It makes consumers feel a whole lot better when a business just plain says something. You're right though, if they do say something and then another month goes by, it's more hell. Damned if you do and damned if you don't. My view though is that they could have minimally just said something like "boost bug is acknowledged, rest assured this is a top priority for our engineers, blahblah...."
 

SilverSliver

Beat It To Deformation
Joined
Feb 23, 2004
Messages
11,230
Congrats you have failed in the first place. For the millionth time max boost is frickin single core boost not multicore. I can not believe how misinformed some here are. AMD has never stated in any of their ryzen processors that max boost is all core. It has always been single core.
I know that, just as much as I know that this is not a 25mhz issue. 4.6ghz boost should mean single core boost under good thermal conditions in single threaded applications, ie - games. I figured it would behave like previous gen Ryzen, and you'd get a multicore overclock into the 4.5 range or you'd see 4.6, occasionally, while gaming. 4.5 all core is a pipedream, and gaming speeds are 4.3ghz - and I'm not even touching the PBO video from AMD.

My e8400, 2600k, 6700k, q6600 - all were able to achieve maximum advertised single core boost in games - and then some with OC. 3900x cannot.

I don't play nop loops, Cinebench or desktop idle states, and those should not be the sole metrics of 'see, it hit max boost!'
 

Gideon

2[H]4U
Joined
Apr 13, 2006
Messages
2,467
Just because this isn't on the same level as Soundblaster driver issues on a 486 doesn't mean it's not a problem. My patience for computer hardware to just work as advertised has been molded over several generations of products now. If, at any point, your customer is wondering if they need to RMA what they have because they cannot figure out if it's a problem with the product or just working as designed, the company has failed in one way or another.
My 3900x hits max boost on one core. Light muliti core loading of any type results in 4.25 to 4.3ghz boost. Heavy multi core is 4.05 to 4.1 boost. For gaming, Intel still has a substantial lead in pure single core and lightly threaded multi core gaming.
Yet here you report normal operation of the chip you own. Higher clocks will depend on keeping the chip cool with more cores used. Just seems like your upset because you think Intel has a big lead yet every reviewer showed much lower clocks with all core use, you seem to not understand all reviews showed the speeds to go down the more you load the chip and quite quickly. Your expectation was never inline with the reality of reviews.
 

NKD

[H]ardness Supreme
Joined
Aug 26, 2007
Messages
7,785
I know that, just as much as I know that this is not a 25mhz issue. 4.6ghz boost should mean single core boost under good thermal conditions in single threaded applications, ie - games. I figured it would behave like previous gen Ryzen, and you'd get a multicore overclock into the 4.5 range or you'd see 4.6, occasionally, while gaming. 4.5 all core is a pipedream, and gaming speeds are 4.3ghz - and I'm not even touching the PBO video from AMD.

My e8400, 2600k, 6700k, q6600 - all were able to achieve maximum advertised single core boost in games - and then some with OC. 3900x cannot.

I don't play nop loops, Cinebench or desktop idle states, and those should not be the sole metrics of 'see, it hit max boost!'
but most games are multi threaded now so I don’t see how you can get max single core boost when most games use multiple cores. Also comparing it to intel is not fair since it has been the case. AMD uses their boost clocks different then intel. Hey maybe they can relax it a bit and allow more boost at higher temps. After all AMD might be doing it based on their stock cooler while intel doesn’t really include any cooler with their CPU. May be AMD should do the same. Allow for higher temps and higher boost.
 

odditory

[H]ardness Supreme
Joined
Dec 23, 2007
Messages
5,867
People gotta have something to complain about..:eek:
That's what we call in the biz... a

upload_2019-9-3_12-41-23.png


I mean where does it stop - you can just apply that to everything when a manufacturer doesn't live up to their promises.

Imagine Tesla advertises a car with 250 mile range, and it releases with 250 mile range, but then they put out a stealth firmware update to scale it back to 225 miles since they realize it could kill the battery quicker. Fair enough. BUT then when people complain they're only getting 225 miles instead of the advertised 250, the response is "Hey man people just gotta have something to complain about".
 

Gideon

2[H]4U
Joined
Apr 13, 2006
Messages
2,467
That's what we call in the biz... a

View attachment 184864

I mean where does it stop - you can just apply that to everything when a manufacturer doesn't live up to their promises.

Imagine Tesla advertises a car with 250 mile range, and it releases with 250 mile range, but then they put out a stealth firmware update to scale it back to 225 miles since they realize it could kill the battery quicker. Fair enough. BUT then when people complain they're only getting 225 miles instead of the advertised 250, the response is "Hey man people just gotta have something to complain about".
So then are you complaining to Nvidia about your lack of DLSS support or the fact that DLSS x2 never arrived? You have a obvious double standard yet have no intention of ever owning the hardware your in here harping about. Also your example is horrible, no manufacturer ever guarantees mileage we give you a window much like AMD did.
 

ccityinstaller

[H]ardness Supreme
Joined
Feb 23, 2007
Messages
4,236
That's what we call in the biz... a

View attachment 184864

I mean where does it stop - you can just apply that to everything when a manufacturer doesn't live up to their promises.

Imagine Tesla advertises a car with 250 mile range, and it releases with 250 mile range, but then they put out a stealth firmware update to scale it back to 225 miles since they realize it could kill the battery quicker. Fair enough. BUT then when people complain they're only getting 225 miles instead of the advertised 250, the response is "Hey man people just gotta have something to complain about".
Do you actually own a a Ryzen 3000 series? How about an AMD product period? You are quick to pop up in every AMD thread just to declare how much better Nvidia and Intel are vs the AMD offerings. You offer NOTHING meaningful to the discussion, and it is frankly tiring. I
n
I'm not there saying there is not a technical reason to complain (from a black/white perspective) but the issue is that things are NOT black/white. The last AGESA code IMPROVED performance, despite the clocks not increasing. That means that you get more performance for your money then what you thought you were getting vs launch day. That is a nice bonus and anyone that doesn't take the time to realize this is foolish.
 

thesmokingman

[H]ardness Supreme
Joined
Nov 22, 2008
Messages
5,699
That's what we call in the biz... a

View attachment 184864

I mean where does it stop - you can just apply that to everything when a manufacturer doesn't live up to their promises.

Imagine Tesla advertises a car with 250 mile range, and it releases with 250 mile range, but then they put out a stealth firmware update to scale it back to 225 miles since they realize it could kill the battery quicker. Fair enough. BUT then when people complain they're only getting 225 miles instead of the advertised 250, the response is "Hey man people just gotta have something to complain about".
Lmao. Tesla actually does. They change the specs at will via OTA updates. Hahaha! Cars have lost range and gained range... they take away features, then add them back etc etc.
 

SilverSliver

Beat It To Deformation
Joined
Feb 23, 2004
Messages
11,230
Yet here you report normal operation of the chip you own. Higher clocks will depend on keeping the chip cool with more cores used. Just seems like your upset because you think Intel has a big lead yet every reviewer showed much lower clocks with all core use, you seem to not understand all reviews showed the speeds to go down the more you load the chip and quite quickly. Your expectation was never inline with the reality of reviews.
I'm at high 60s during typical loads. How cool does my chip need to be? 50C? 40C? 30C? Do I need to dig up a Primochill to get box speeds?

I'm irritated because I spent a month with a chip that wouldn't go past 4325 at all - working with Gigabyte staff, extensive testing and compiling for the community to try and figure out the problem - all the while folks like you and other fanboys heaped all the problems on me and the others having the issues. ABB bios comes out and, magically, fixes most of the boost problems. Now, the problem is being reported as still existing, in various forms, and more widespread than typically thought and all the same people are out defending AMDs honor - blaming it on everything from case fans to dust bunnies as to why AMD cannot deliver real single core boost clocks anywhere near what they advertised and why PBO doesn't work at all.

Advertising a max boost, and that max boost only happening on a generation old Cinebench tool , idle state or nop loops while using a specific polling rate in a specific hardware monitoring software suite as dictated by AMD is 100% disingenuous to the reality of use. This, coupled with the PBO fiasco - to me - indicates absolute and intentional deception from AMD marketing for this series of chips.
 

Master_shake_

[H]ardForum Junkie
Joined
Apr 9, 2012
Messages
10,682
I would spontaneously combust in those temperatures. I would keep it colder if I could. My girlfriend is the one that puts the thermostat at 73F. I'd run it at 70F-71F if I had it my way.
letting the gf touch the thermostat.

that's your first problem.
 
Joined
Mar 23, 2013
Messages
721
I would spontaneously combust in those temperatures. I would keep it colder if I could. My girlfriend is the one that puts the thermostat at 73F. I'd run it at 70F-71F if I had it my way.
69°F during the summer months here, similar in the winter. The wife and I don't really spend a lot on ourselves (going out to eat, vaca's, etc) so spending a bit more on cooling/electricity in the summer is our gift to ourselves, I guess you could say :LOL:

letting the gf touch the thermostat.

that's your first problem.
QFT :ROFLMAO:

What you do is make the public thermostat a dummy one, don't hook it up. Then put a second thermostat where only you know. That way the wife gets to "control the temp" but in reality, you'll know better. Saving marriages, it's what I do. (y)
 

reaper12

2[H]4U
Joined
Oct 21, 2006
Messages
2,306
What a thread. This is what passes as a debacle these days?

Some of you guys need to get out more.
 

ChadD

[H]ardness Supreme
Joined
Feb 8, 2016
Messages
4,518
Imagine Tesla advertises a car with 250 mile range, and it releases with 250 mile range, but then they put out a stealth firmware update to scale it back to 225 miles since they realize it could kill the battery quicker. Fair enough. BUT then when people complain they're only getting 225 miles instead of the advertised 250, the response is "Hey man people just gotta have something to complain about".
I hope I'm not being the jerk telling odditory that Santa doesn't exist.

Car Manufacturers do exactly that ALL the time. No firmware updates needed to complicate reality. Unless you are the one customer in the world that believes the sticker on the side of your new wheels that says you will get X miles per gallon. Ya manufacturers advertise best case scenarios numbers... that isn't a new thing. Same goes for things like HP ratings... if you believe your brand new vehicle off the showroom floor hits its rated HP you would be most likely wrong. A few might most won't. Perfect set of new tires, perfect stretch of road at the exact right acceleration curves and ya you might match the rated fuel economy. No real difference here... perfect setup perfect board delivering the perfect amount of voltage on the perfect curve with the right ram and Inf fabric clocks ect your chip can hit Max single core boost.

Having said that of course we know they are releasing an updated AGEIS shortly that may push more people to or over MAX boost. Guess we'll see in a few weeks.
 

Gideon

2[H]4U
Joined
Apr 13, 2006
Messages
2,467
I'm at high 60s during typical loads. How cool does my chip need to be? 50C? 40C? 30C? Do I need to dig up a Primochill to get box speeds?

I'm irritated because I spent a month with a chip that wouldn't go past 4325 at all - working with Gigabyte staff, extensive testing and compiling for the community to try and figure out the problem - all the while folks like you and other fanboys heaped all the problems on me and the others having the issues. ABB bios comes out and, magically, fixes most of the boost problems. Now, the problem is being reported as still existing, in various forms, and more widespread than typically thought and all the same people are out defending AMDs honor - blaming it on everything from case fans to dust bunnies as to why AMD cannot deliver real single core boost clocks anywhere near what they advertised and why PBO doesn't work at all.

Advertising a max boost, and that max boost only happening on a generation old Cinebench tool , idle state or nop loops while using a specific polling rate in a specific hardware monitoring software suite as dictated by AMD is 100% disingenuous to the reality of use. This, coupled with the PBO fiasco - to me - indicates absolute and intentional deception from AMD marketing for this series of chips.
Here is all core boost chart by temp. They didn't do single core tests but assume it will scale by temperature as well. No one deceived you if you actually read reviews, it was quite obvious what these chips did.

amd-ryzen-3900x_cold-scale_1.png
 

Dayaks

[H]ardness Supreme
Joined
Feb 22, 2012
Messages
7,531
Here is all core boost chart by temp. They didn't do single core tests but assume it will scale by temperature as well. No one deceived you if you actually read reviews, it was quite obvious what these chips did.

View attachment 184900
You mean, what they couldn’t do. ;)

I feel similar to this as I did the GTX 970. Seems shitty but I read reviews before I buy things. I do understand how someone would be pissed though. Buuttt nVidia did get sued...
 

sabrewolf732

[H]ardness Supreme
Joined
Dec 6, 2004
Messages
4,152
That's what we call in the biz... a

View attachment 184864

I mean where does it stop - you can just apply that to everything when a manufacturer doesn't live up to their promises.

Imagine Tesla advertises a car with 250 mile range, and it releases with 250 mile range, but then they put out a stealth firmware update to scale it back to 225 miles since they realize it could kill the battery quicker. Fair enough. BUT then when people complain they're only getting 225 miles instead of the advertised 250, the response is "Hey man people just gotta have something to complain about".
I think a more realistic comparisson would be manufacturer HP ratings, when tested HP ratings often don't match up as the manufacturer frequently quotes crank horsepower. What does matter ultimately is the numbers the car puts down on a track, and ultimately ryzen 3000 is putting down very good numbers at the track. I think the boost issue is a genuine issue and is dishonest, AMD definitely knew about it.

Does it make ryzen 3000 bad? No.
 

Ocellaris

Ginger @le, an alcoholic's best friend.
Joined
Jan 1, 2008
Messages
18,848
I think a more realistic comparisson would be manufacturer HP ratings, when tested HP ratings often don't match up as the manufacturer frequently quotes crank horsepower. What does matter ultimately is the numbers the car puts down on a track, and ultimately ryzen 3000 is putting down very good numbers at the track. I think the boost issue is a genuine issue and is dishonest, AMD definitely knew about it.

Does it make ryzen 3000 bad? No.
Well they definitely know about it now...
 

Jim Kim

2[H]4U
Joined
May 24, 2012
Messages
3,646
I'm at high 60s during typical loads. How cool does my chip need to be? 50C? 40C? 30C? Do I need to dig up a Primochill to get box speeds?

I'm irritated because I spent a month with a chip that wouldn't go past 4325 at all - working with Gigabyte staff, extensive testing and compiling for the community to try and figure out the problem - all the while folks like you and other fanboys heaped all the problems on me and the others having the issues. ABB bios comes out and, magically, fixes most of the boost problems. Now, the problem is being reported as still existing, in various forms, and more widespread than typically thought and all the same people are out defending AMDs honor - blaming it on everything from case fans to dust bunnies as to why AMD cannot deliver real single core boost clocks anywhere near what they advertised and why PBO doesn't work at all.

Advertising a max boost, and that max boost only happening on a generation old Cinebench tool , idle state or nop loops while using a specific polling rate in a specific hardware monitoring software suite as dictated by AMD is 100% disingenuous to the reality of use. This, coupled with the PBO fiasco - to me - indicates absolute and intentional deception from AMD marketing for this series of chips.
Did you buy the chip to stare at the frequency, because if that's why you bought it than by all means request a refund. But if you bought it to use, use it.
 

Revdarian

2[H]4U
Joined
Aug 16, 2010
Messages
2,518
That’s awesome you guys turned 10%+ less performance into some kind of weird generational thing. I don’t even know how you guys managed that.

So the narative is now you’re a soy boy if you don’t let AMD gape you. Seems kind of backwards?
It is not 10%, it isn't even 5% as that would see 4180 MHz instead of 4400 MHz, and the same agesa actually increased effective single thread performance shown on anandtech, in any case amd will release that microcode with the boost, it may not achieve performance parity with the current release, could even have a small regression, we shall see once it is done.


That's what we call in the biz... a

View attachment 184864

I mean where does it stop - you can just apply that to everything when a manufacturer doesn't live up to their promises.

Imagine Tesla advertises a car with 250 mile range, and it releases with 250 mile range, but then they put out a stealth firmware update to scale it back to 225 miles since they realize it could kill the battery quicker. Fair enough. BUT then when people complain they're only getting 225 miles instead of the advertised 250, the response is "Hey man people just gotta have something to complain about".
Entirely wrong car analogy, proper would be relating the hertz to rpm since like anandtech showed, the agesa post release, the main culprit lowering max single thread boost, actually increased single thread performance.

So, a type of microcode change to your tesla lowered the max rpm, it didn't touch the acceleration, it actually increased the max effective speed, and so far no change has been seen on the fuel efficiency.
 

Dayaks

[H]ardness Supreme
Joined
Feb 22, 2012
Messages
7,531
It is not 10%, it isn't even 5% as that would see 4180 MHz instead of 4400 MHz, and the same agesa actually increased effective single thread performance shown on anandtech, in any case amd will release that microcode with the boost, it may not achieve performance parity with the current release, could even have a small regression, we shall see once it is done.




Entirely wrong car analogy, proper would be relating the hertz to rpm since like anandtech showed, the agesa post release, the main culprit lowering max single thread boost, actually increased single thread performance.

So, a type of microcode change to your tesla lowered the max rpm, it didn't touch the acceleration, it actually increased the max effective speed, and so far no change has been seen on the fuel efficiency.
I was going off what the AMD marketing team informed the community days before launch.
 
Top