Support NeoGAF

[cebri.one]

So who is tidux?

PS4 dev apparently

 

[Vizzeh]

So who is tidux?

Click on his quote and read through that thread he is in, he is very active in it, seems to be a developer of a game working on the snowdrop engine, possibly the division, or he is being cryptic about another game, interesting thread and insight tbh, he does mention the PS4 is much faster/easier etc.

He also says Ubisoft about watchdogs is only the first of few top devs to admit the PS4 advantage. (more will come)

 

[Chobel]

PS4 dev apparently

Click on his quote and read through that thread he is in, he is very active in it, seems to be a developer of a game working on the snowdrop engine, possibly the division, or he is being cryptic about another game, interesting thread and insight tbh, he does mention the PS4 is much faster/easier etc.

He also says Ubisoft about watchdogs is only the first of few top devs to admit the PS4 advantage.

Thanks, was just making sure he legitimate source.

 

[Manmademan]

Thanks, was just making sure he legitimate source.

sources that claim to be legitimate developers or insiders get banned VERY quickly if they can't back it up to moderators. No need to worry about that one here, at least.

 

[shinra-bansho]

Has he been mod-verified, and wouldn't a dev discussing some of this stuff be NDA'd against doing so?

 

[Skeff]

sources that claim to be legitimate developers or insiders get banned VERY quickly if they can't back it up to moderators. No need to worry about that one here, at least.

Yea, I haven't read a mod confirmation, but he's been dishing info out like candy on Halloween, so I'm sure miss will have verified him.

 

[Vizzeh]

Thanks, was just making sure he legitimate source.

seems like good info from here:

http://www.neogaf.com/forum/showthread.php?t=686899&page=8

 

[Perkel]

B3D is surely reaching banned site status?

The abdication of modding responsibilities at that site has seen it become a cesspool of FUD. It really has become what AVS was during the HD format wars.

B3D is still fantastic site to talk about tech.

Just because few juniors quote same few people Who type there to boost their credits we know from B3D xbox thread it doesn't mean that B3D is shitty site.

I value shifty geezer word much more than almost anyone from Neogaf.

 

[Lactose_Intolerant]

B3D is surely reaching banned site status?

The abdication of modding responsibilities at that site has seen it become a cesspool of FUD. It really has become what AVS was during the HD format wars.

It is sad because it was a great site once. I haven't gone there in a while. I think I stopped when Dave got hired by ATI/AMD.

 

[stryke]

He seems to be an okay source, but doesn't seem to be very subtle. Especially when we've already managed to get a name and face to Tidux, not sure what his employers think of him mentioning these sort of things. Alas, he's lasted this long, don't think statements like his would be missed by mods.

Demonite on the other hand, it's quite rare for him to actually comment console performance. Nice to see him pitch in.

 

[Vizzeh]

He seems to be an okay source, but doesn't seem to be very subtle. Especially when we've already managed to get a name and face to Tidux, not sure what his employers think of him mentioning these sort of things. Alas, he's lasted this long, don't think statements like his would be missed by mods.

Demonite on the other hand, it's quite rare for him to actually comment console performance. Nice to see him pitch in.

Dont think he is disclosing any information we dont already know, just seems to be commenting on public knowledge really but its good to hear it from a dev and his opinion, he was quite cagey about the stuff he isnt supposed to mention

 

[Jack_AG]

B3D is an excellent source for laughs these days. Glad to see juniors linking such excellent comedy here. Saves me the work of actually sifting through B3D's bullshit to get to the fun stuff

On that note, we need to pack shit up. No amount of developer commentary citing personal experiences developing on both platforms will ever quell the uprising of "junior"s spouting nonsense and spreading FUD.

Sorry, juniors, I'll take a dev's word over yours all day every day - even if its news I don't want to hear.

 

[shinra-bansho]

He seems to be an okay source, but doesn't seem to be very subtle. Especially when we've already managed to get a name and face to Tidux, not sure what his employers think of him mentioning these sort of things. Alas, he's lasted this long, don't think statements like his would be missed by mods.

Demonite on the other hand, it's quite rare for him to actually comment console performance. Nice to see him pitch in.

The new guy, assuming he isn't bishslapped presumably works for Ubi considering he talks about the Snowdrop engine.

Also someone found his home address, lol. DetectiveGAF is scary GAF.

What does DemonNite do/who does he work for btw - such that he's an insider?

 

[fastford58]

From this thread

Not that we need more evidence, but more is always nice.

 

[haterofpants]

Xbox One Retail Version of Battlefield 4 Will Still Run at 720p, 60 FPS, EA Representative Confirms. PS4 version will run at a higher resolution, I'm assuming at least 900p like at TGS.

Both the Xbox One and PS4 version of Battlefield 4 were demonstrated running at 720p resolution and 60 FPS, but as confirmed by Frostbite Technical Director Johan Andersson on Twitter, it runs at a higher resolution in the PS4 developer environment and the final resolution will be apparent when the game will launch.

Can Xbox One users planning to get the game hope for the same treatment? Apparently not, as confirmed by an “EA representative in charge” to the Japanese site Game*Spark, that got hands on time with the game at Eurogamer Expo.

According to said EA rep, the Xbox One retail version of the game will still run at 720p , 60 fps, the same as the alpha version showcased at Eurogamer Expo.

In addition to that Game*Spark’s writer mentions that graphics “look blurred and a little flat, probably due to lower resolution textures and somewhat worse lighting compared to the PC version”.

 

[miDnIghtEr20C]

So the people that talk about the ESRAM all the time like people who talked about the CELL all the time? I'm not a huge tech guy but is that close to the same style? Because I remember a ton of talk about the CELL for a few Years when the PS3 launched.

 

[TheRealTalker]

The new guy, assuming he isn't bishslapped presumably works for Ubi considering he talks about the Snowdrop engine.

Also someone found his home address, lol. DetectiveGAF is scary GAF.

What does DemonNite do/who does he work for btw - such that he's an insider?

I believe he is a Sony Employee (might be wrong)

 

[stryke]

I believe he is a Sony Employee (might be wrong)

Correct.

 

[Jack_AG]

Xbox One Retail Version of Battlefield 4 Will Still Run at 720p, 60 FPS, EA Representative Confirms. PS4 version will run at a higher resolution, I'm assuming at least 900p like at TGS.

So the slugline is "confirmed" but later in the article DS says to take it with a grain of salt and that he contacted EA for further confirmation but got no reply... ... ... I love how "journalists" always try to pass off news leaving themselves an out as thinking somehow it absolves them from any misinformation. I mean... I'm all for a better version on PS4 but that whole article smells fishy.

 

[shinra-bansho]

So the people that talk about the ESRAM all the time like people who talked about the CELL all the time?

No.

Properly utilizing the Cell to make up for the poor RSX and more difficult memory set-up on the PS3, developers could match or exceed the overall performance of the 360, with it's better GPU.

The ESRAM essentially provides a fast but tiny scratchpad to compensate for the large pool of slow main RAM. Under certain circumstances you may be able to, for 32MB of data achieve the same or faster bandwidth than the PS4. But the PS4 has 8GB of RAM all operating at a fast bandwidth.

The ESRAM doesn't do anything to change the difference in CUs, TMUs and ROPs.

That doesn't mean that XB1 games aren't going to look great. That doesn't mean the difference is necessarily going to be staggering between the two, it remains to be seen how much the theoretical deltas translate to real world differences. But it's a false analogy.

 

[Wynnebeck]

Xbox One Retail Version of Battlefield 4 Will Still Run at 720p, 60 FPS, EA Representative Confirms. PS4 version will run at a higher resolution, I'm assuming at least 900p like at TGS.

It begins.

 

[ebullientprism]

B3D is an excellent source for laughs these days. Glad to see juniors linking such excellent comedy here. Saves me the work of actually sifting through B3D's bullshit to get to the fun stuff

On that note, we need to pack shit up. No amount of developer commentary citing personal experiences developing on both platforms will ever quell the uprising of "junior"s spouting nonsense and spreading FUD.

Sorry, juniors, I'll take a dev's word over yours all day every day - even if its news I don't want to hear.

You are so cool. I bet all those "juniors" wish they had more posts like you do. They all must be so jealous of you.

 

[OldSchoolNerd]

DF article is crap, too.
Thanks to the small clock boost, etc.... and speaking about AAA games -->around ~140GB/s is what the XBone will achieve, if everthing goes as it should.
PS4 is currently at ~172GB/s.
edit.

No, according to the DF article and what microsoft says they get 200GB/s aggregate bandwidth.
Now if you continue to state that the x1 only has 140GB/s then you are saying MS/DF are lying. Which is a view that you or anybody is entitled to, but its a bit tinfoil hat to me.

Direct quote from the article.

"The same discussion with ESRAM as well - the 204GB/s number that was presented at Hot Chips is taking known limitations of the logic around the ESRAM into account. You can't sustain writes for absolutely every single cycle. The writes is known to insert a bubble [a dead cycle] occasionally... one out of every eight cycles is a bubble so that's how you get the combined 204GB/s as the raw peak that we can really achieve over the ESRAM. And then if you say what can you achieve out of an application - we've measured about 140-150GB/s for ESRAM.

"That's real code running. That's not some diagnostic or some simulation case or something like that. That is real code that is running at that bandwidth. You can add that to the external memory and say that that probably achieves in similar conditions 50-55GB/s and add those two together you're getting in the order of 200GB/s across the main memory and internally."

So 140GB-150GB is a realistic target and DDR3 bandwidth can really be added on top?

"Yes. That's been measured."

 

[OldSchoolNerd]

Neither L1 nor L2 are system memory, and their latency is in the 100s of cycles. Is that "extremely low"?

Yes. Its lower than system ram. Not really sure I get the point you are trying to make?

 

[Jack_AG]

You are so cool. I bet all those "juniors" wish they had more posts like you do. They all must be so jealous of you.

OK!

 

[Manmademan]

No, according to the DF article and what microsoft says they get 200GB/s aggregate bandwidth.
Now if you continue to state that the x1 only has 140GB/s then you are saying MS/DF are lying. Which is a view that you or anybody is entitled to, but its a bit tinfoil hat to me.

Direct quote from the article.

Not that this isn't very entertaining, but you're aware that junior bans are permanent, right? Just so we're clear?

 

[OldSchoolNerd]

Oh come on. This bullshit has been called out countless time on this very thread. Are we gonna go all over it again?
Bandwidth doesn't add up. It doesn't. If it did, x360 would have more than x1.
Peak performance was measured at 133GB/s before the upclock. Now you can assume it's 140. And no, 3/4 of ram requests aren't coming off the eSram, and yes, everything that has to make a journey in DDR3 is limited to 68GB/sec theoretically.

Well acording to MS it does add up.

"That's real code running. That's not some diagnostic or some simulation case or something like that. That is real code that is running at that bandwidth. You can add that to the external memory and say that that probably achieves in similar conditions 50-55GB/s and add those two together you're getting in the order of 200GB/s across the main memory and internally."

So 140GB-150GB is a realistic target and DDR3 bandwidth can really be added on top?

"Yes. That's been measured."

3/4 or ram requests coming of esram?

We have 150GB/s from esram and 50GB/s from ddr3. Each second, over all the frames they are seeing 3/4 of the requests server from esram. Ie 150/200 = 3/4.

 

[Abriael]

So the slugline is "confirmed" but later in the article DS says to take it with a grain of salt and that he contacted EA for further confirmation but got no reply... ... ... I love how "journalists" always try to pass off news leaving themselves an out as thinking somehow it absolves them from any misinformation. I mean... I'm all for a better version on PS4 but that whole article smells fishy.

Way to misinterpret it. The rep indeed confirmed it, exactly as the headline says.

But since it's a rep, I reached out to EA out of scruple. the "Got no reply" part is also inaccurate, as of course they didn't reply YET. I mailed them two hours ago on a sunday.

 

[OldSchoolNerd]

We've been over this many, many times. Latency is barely even a factor for GPU tasks, bandwidth is what's important for that.

I agree GPU tasks are latency tollerant, however what I am suggesting is that if you actually had very low latency ram it would allow the GPUs to get to higher utilisation....but this is not gospel. Just something that makes sense to me in trying to work out how MS call their system balanced with 200GB/s going to 12CUs. It only makes sense if they are getting those CUs up to higher levels of utilsation that could be achived with a higher latency ram.

 

[Jack_AG]

Way to misinterpret it. The rep indeed confirmed it, exactly as the headline says.

But since it's a rep, I reached out to EA out of scruple. the "Got no reply" part is also inaccurate, as of course they didn't reply YET. I mailed them two hours ago on a sunday.

The rep confirmed it - but then you said to take all that information with a "bit of a grain of salt" - so I did EXACTLY as you instructed and took it with a grain of salt - if you did not want that to happen - perhaps you shouldn't tell people to take things with a grain of salt. Also - if you didn't get an email reply, then I am correct in saying you "got no reply". "Yet" is not an indicator or timestamp at the time of writing/publishing.

 

[AgentP]

Well acording to MS it does add up.



3/4 or ram requests coming of esram?

We have 150GB/s from esram and 50GB/s from ddr3. Each second, over all the frames they are seeing 3/4 of the requests server from esram. Ie 150/200 = 3/4.

You are either Rangers or enjoy using his crap, not sure which is worse.

 

[Chobel]

Well acording to MS it does add up.



3/4 or ram requests coming of esram?

We have 150GB/s from esram and 50GB/s from ddr3. Each second, over all the frames they are seeing 3/4 of the requests server from esram. Ie 150/200 = 3/4.

No, that's not how it works. Ram bandwidth have nothing to do number of requests...
Also we don't know what kind of scenarios in which this bandwidth happens or how common it is.

 

[Insane Metal]

Well acording to MS it does add up.



3/4 or ram requests coming of esram?

We have 150GB/s from esram and 50GB/s from ddr3. Each second, over all the frames they are seeing 3/4 of the requests server from esram. Ie 150/200 = 3/4.

Uh...

 

[Respawn]

GAF will always be more accurate and reliable, because here we have a strict control and check up system where insiders etc are vetted before they're allowed to spout BS claims. After the sifting through the weed, this generally leaves a better cut of actual insiders and industry people left posting. Many of the ones who've posted in here even, despite not being insiders, still have experience in the field or fields similar, and can usually pretty quickly and astutely shoot down FUD. It's a shame so much FUD on B3D is not only propagated, but actually regurgitated and continually given traction.

The majority of secret sauce developments actually originated from B3D, that should tell you what you need to know credibility wise. Luckily they do still have a few genuine insiders and devs etc who do still post, albeit not nearly enough.

I've been a member there since 05 I think but the site went to shit a long time ago. Last time I posted there was when Reiko got banned. I remember the good ol days when PS3 wasn't launched yet and Ninja Theory devs use to post there giving some insight on Heavenly Sword. These days it's comical to see the stuff on there.

 

[spatenfloot]

Not that this isn't very entertaining, but you're aware that junior bans are permanent, right? Just so we're clear?

Don't worry. There will be a new one to replace him when that happens.

 

[_Ryo_]

Using "We" seems to be a weird way to phrase something unless you're affiliated with the source (MS). IF you're affiliated with the source and knowingly spewing nonfactual information and PR, well, good luck.

 

[OldSchoolNerd]

Are you referring to the zupallinere post? The one where he/she quotes you, and makes the assumption the xbox one is a monster on the back of your comment? Wow.

Ha! No, but I did like his phrasing! I was refering to the other senior guys on there that agree with what I am trying to get across. For example.

Shifty Geezer, member since dec 2004 with 27k posts:

Tackling people who doubt what the 150GB/s esram bandwidth means

Why is 'we measured' not good enough? That's actually the best way to ascertain real-world performance; actually use the design in the real world and see how it performs. 150 GB/s should mean they shifted 150 GB within the span of one second. That may have been sustained 150 GB/s for a second, or a mix for 30 GB/s and 200 GB/s over that second.

I don't understand why people are arguing so much over these BW figures. We have an ESRAM bus structure, and explanation why it isn't read+write simultaneous 100% of the time, clockspeeds and bus widths, and finally some measured performance from real-world use that seems eminently plausible, favouring the middle of the road between high and low. We can't get any more info than that unless we get hold of game performance profiles from the dev tools. Maybe with XB1 being an SDK, someone will manage to do that somehow. Until then, this topic seems kinda done to death.

Talking about esram bandwidth

As for the conversation, I'll post a summary that'll hopefully bring it to a natural close instead of everyone delving into cyclic arguments to nowhere. Bandwidth is a measure of data capacity over time. Whenever measured, it's an average (even a peak BW measurement is an average over time). That MS measured 150 GB/s means they achieved 150 GB/s. That's with considerable probability. There's no reason whatsoever for anyone to believe that was 150 GB/s for a fraction of a second and the average BW for ESRAM access is notably less than that; that defies the purpose of the measurement as an average and a metric you'd use to inform your development partners. That sort of argument is basically looking for an opportunity to doubt the evidence in front of is, and is unnecessary.

Complaining about people trying to judge the performance of the system with a single number....

Aaaaaaa! It is not possible to understand the flow of data in a system by a single metric (unless that system has a single memory pool). Your aggregate number is true and yet pointless, and there's zero sense in trying to condense understanding of the BW into this single value.

Sometimes the code will run at the fastest aggregate speed of the total RAM. Sometimes it could be bottlenecked by the slowest singular pipe. Mostly it'll be hitting shifting limits as data moves around the different pools. All games will have access to ~200 GB/s (actually 272 GB/s as total peak available BW) but the amount of data flowing through the system could be very different. The most important thing is that devs will try to maximise dataflow within budgets and development targets, which is why they want to know bus speeds. Bus speeds aren't really for informing the masses about the potential of the consoles!

Gubbi, member since feb 2002, 2700 posts.

The read to write ratio is something like 2-4:1 for normal graphics workloads. If we assume we use all of the ESRAM read bandwidth (ie >95%) and 60 % of the DDR3 bandwidth for GPU reads, we get around 45-50GB write traffic to ESRAM with a 3:1 ratio.

That's 150GB/s (100 read, 50 write) of ESRAM bandwidth with aggregate system bandwidth being well over 200GB/s. Lower average latency too.

Cheers

Originally Posted by MrFox View Post
They discovered that it was designed with a full bidirectional interface. Someone somehow forgot to tell anyone.

Some pointy hair bosses, marketing drones, and a bunch of internet tards suddenly discovered the aggregate bandwidth numbers.

The engineers who designed the thing didn't do it based on assumptions and luck. They examined various workloads and existing bottlenecks, designed the thing to have separate read and write buses to/from the ESRAM knowing all along what typical aggregate bandwidth could be expected for various graphics engine implementations.

As did the ones working on PS4.

MS ended up with more aggregate bandwidth and lower average latency. Sony ended up with a lot more computational grunt. That's because they looked at different workloads and designed for different price points.

Cheers

Silent_Buddah, member since Mar 2007, 9906 posts

Aaaaaaa! It is not possible to understand the flow of data in a system by a single metric (unless that system has a single memory pool). Your aggregate number is true and yet pointless, and there's zero sense in trying to condense understanding of the BW into this single value.

Sometimes the code will run at the fastest aggregate speed of the total RAM. Sometimes it could be bottlenecked by the slowest singular pipe. Mostly it'll be hitting shifting limits as data moves around the different pools. All games will have access to ~200 GB/s (actually 272 GB/s as total peak available BW) but the amount of data flowing through the system could be very different. The most important thing is that devs will try to maximise dataflow within budgets and development targets, which is why they want to know bus speeds. Bus speeds aren't really for informing the masses about the potential of the consoles!

Now you can say what you like about what I know etc, but I am quoting information from MS and then the views from the guys at beyond3d. If you think these guys are full of it then thats up to you. Having read a lot of stuff on their site, I believe them.

 

[AgentP]

Keep quoting those die hard b3d xbots, especially the non Dev forum warriors... Not helping your case.

 

[Bundy]

OldSchoolNerd..... stahp!

 

[OldSchoolNerd]

Bullshit.
The real world practical bandwidth numbers will likely be:

XB1: 130-140 GB/s
PS4: 172 GB/s

The PS4 number has already been confirmed by a developer's CEO on this forum.

The esram can only provide speeds of 150gb/s maximum according to MS at the moment and that is pushing it to specifically do only certain tasks and with it being as full as possible of data for those tasks. Please don't spout Beyond3D bullshit here on GAF as that shit get's called out real fast.

EDIT: btw the xb1 number includes the DDR3 bandwidth too, the fact the Esram is not a 1024bit bus and is actually 4x 256 bit bus attached to 4 8mb sections means it will generally be operating at far lower than any of these theoretical numbers in a real world scenario.

Actually the way I see it looks like this.

X1, peak 274GB/s average 200GB/s
PS4, peak 176GB/s average 130GB/s

Now the only figure that can possibly be in dispute is the PS4 average figure as Sony havent been forthcoming with any test results. But MS state that

However just like main memory, it's rare to be able to achieve that over long periods of time so typically an external memory interface you run at 70-80 per cent efficiency.

If you take the middle of that its 75% hence 130GB/s. Now I have heard guys going on about "The ps4 gets 172GB/s" which has its source in a web site where there is a quote from a director of an indi games company. No unless somehow sony have silently solved the intractable average vs peak bandwidth problem that has been in existence since the year dot, I would suggest thats a mis-quote/mis-speak etc and he was actually referring to the peak bandwidth of 176GB/s.

Maybe its 140, or if they actually managed to achieve 85% they could get 150, but even that is a long way south of 200.

This article goes into depth as to why http://archive.arstechnica.com/paedi...latency-1.html you never achieve peak for any great amount of time.

With regards to they beyond3d bullshit...hummm to be honest I think they hold their posters to pretty high standards. They certainly get well moody when you try and make a direct comparison between the x1 and the ps4!

 

[Ishida]

Actually the way I see it looks like this.

X1, peak 274GB/s average 200GB/s
PS4, peak 176GB/s average 130GB/s

What is this I don't even

 

[Boombloxer]

What is this I don't even

Magic math that was corrected pages ago.

Quite amazing, really. It's the exact same argument.

 

[NTIETS13A_Superiority]

.

Beyond3d forums has been joke of a site for years now. Most important is what one does whit esram bandwith.. there is only so much one can do with only 32 megs of it.

 

[Cidd]

I see specialguy/Rangers has return...

 

[artist]

What is this I don't even

Senju's backup is here.

 

[Ishida]

Senju's backup is here.

Seriously. Is this the same guy with different new accounts?

(Yay! Not a junior anymore!)

 

[tfur]

What is the read speed from esram?

Ask the technical fellow or Albert to answer that simple question. Not the read and write speed combined, but just the read speed or the write speed.

 

[ekim]

Seriously. Is this the same guy with different new accounts?

(Yay! Not a junior anymore!)

SenjutsuSage isn't banned anymore.

 

[velociraptor]

Actually the way I see it looks like this.

X1, peak 274GB/s average 200GB/s
PS4, peak 176GB/s average 130GB/s

Now the only figure that can possibly be in dispute is the PS4 average figure as Sony havent been forthcoming with any test results. But MS state that



If you take the middle of that its 75% hence 130GB/s. Now I have heard guys going on about "The ps4 gets 172GB/s" which has its source in a web site where there is a quote from a director of an indi games company. No unless somehow sony have silently solved the intractable average vs peak bandwidth problem that has been in existence since the year dot, I would suggest thats a mis-quote/mis-speak etc and he was actually referring to the peak bandwidth of 176GB/s.

Maybe its 140, or if they actually managed to achieve 85% they could get 150, but even that is a long way south of 200.

This article goes into depth as to why http://archive.arstechnica.com/paedi...latency-1.html you never achieve peak for any great amount of time.

With regards to they beyond3d bullshit...hummm to be honest I think they hold their posters to pretty high standards. They certainly get well moody when you try and make a direct comparison between the x1 and the ps4!

You have absolutely no proof to back those figures up. Your post is full of conjecture and speculation.

Also, curious post history. Heh.

 

[OldSchoolNerd]

What is the read speed from esram?

Ask the technical fellow or Albert to answer that simple question. Not the read and write speed combined, but just the read speed or the write speed.

Its 109GB/s. Peak. MS say so in the DF article.

"[ESRAM has four memory controllers and each lane] is 256-bit making up a total of 1024 bits and that in each direction. 1024 bits for write will give you a max of 109GB/s and then there's separate read paths again running at peak would give you 109GB/s.