By using this site, you agree to our Privacy Policy and our Terms of Use. Close

Forums - PC - Carzy Zarx’s PC Gaming Emporium - Catch Up on All the Latest PC Gaming Related News

Jizz_Beard_thePirate said:

Possible AMD RDNA 5 / UDNA GPU SKU Configs Point To 96, 40, 24, 12 CU Dies

https://wccftech.com/possible-amd-rdna-5-udna-gpu-sku-configs-point-to-96-40-24-12-cu-dies/

This just seems to be regurgitating part of MLID leaks. Except they are counting the WGPs as CUs which results the in CU count being halved. It should be 192, 80, 48, and 24. 



Around the Network
Jizz_Beard_thePirate said:

Possible AMD RDNA 5 / UDNA GPU SKU Configs Point To 96, 40, 24, 12 CU Dies

https://wccftech.com/possible-amd-rdna-5-udna-gpu-sku-configs-point-to-96-40-24-12-cu-dies/

&

Darc Requiem said:
Jizz_Beard_thePirate said:

Possible AMD RDNA 5 / UDNA GPU SKU Configs Point To 96, 40, 24, 12 CU Dies

https://wccftech.com/possible-amd-rdna-5-udna-gpu-sku-configs-point-to-96-40-24-12-cu-dies/

This just seems to be regurgitating part of MLID leaks. Except they are counting the WGPs as CUs which results the in CU count being halved. It should be 192, 80, 48, and 24. 

I know it's very early on when it comes to RDNA5/UDNA leaks, but that CU configuration makes me somewhat worried, unless Darc Requiem's point is true.

After all, and assuming that the 96 CU chip will be used for both the 10090 and 10080 GPUs, that leaves us 40 CUs for the 10070 cards when the 9070XT has 64 CUs. Will AMD be able to repeat, for a second time in a row, the feat of increasing the performance per core so much so that it can beat the previous card featuring more cores (the 9070XT beats the 7800XT despite one having 64 and the other 96 CUs)?

But again, if they are actually 80 CUs, that would make more sense. At least, with a 192-bit memory bus size, that means the 10070 could get 24GB of VRAM.

Still, I don't like the big jump between the two top tier chips, with the big one being more than twice the size of the other.

In any case, it's still too soon to know for sure. Things may change once the first samples are made and AMD may make some changes here and there. 

Jizz_Beard_thePirate said:

AMD Reportedly Preparing 16 GB Radeon RX 9070 GRE Graphics Card, Launch Expected In September-October

https://wccftech.com/amd-preps-16-gb-radeon-rx-9070-gre-launch-in-september-october/

I'm having a hard time figuring what AMD is trying to do with this one.

The 9070 GRE and the 9070 non-XT are already quite close in terms of performance, with the memory configuration being their biggest difference so, why change that to make it more complicated? At this point, it would make more sense to just kill the 9070GRE and cut the price of the base 9070... although AMD would need to finally be able to get enough of those chips to the cards could come close to MSRP.



Please excuse my bad English.

Former gaming PC: i5-4670k@stock (for now), 16Gb RAM 1600 MHz and a GTX 1070

Current gaming PC: R5-7600, 32GB RAM 6000MT/s (CL30) and a RX 9060XT 16GB

Steam / Live / NNID : jonxiquet    Add me if you want, but I'm a single player gamer.

I'm not going to pretend to be hyper tech-savy but just on the face of it. The WCCF Tech article doesn't add up. A 12CU dGPU just make zero sense. That's what their mainstream current gen APUs top out at. Their weakest dGPU, the 9060 (non XT), has 28 CUs and that's been their low water mark since RDNA3 (7600S laptop GPU). I know there is a segment of the internet that has a huge hate boner for MLID. I get it, the guy can be smug at times, but he gets far too many things correct to be pulling things out of his rear like the average tech article commenter would have us believe. Given the alleged IPC improvements of RDNA5/UDNA a 24, not 12, CU GPU would have raster performance similar to a 9060 with substantially better RT and PT performance. That's believable.

We'll see the truth of the matter at some point. However the leaked GPU philosophy that AMD is seemingly going for is very smart. It allows them to use the same GPU dies across multiple product segments. The memory solution, especially with the low end SKUs, provides a potential good product for the entry level market. It would finally move us beyond the ridiculous 8GB frame buffer while being cheap enough from a business standpoint to actually come to market. That said, AMD often has good ideas on paper for their GPU business. For the sake of competition, I hope it pans out this time.



The biggest problem is that Kepler, that has a better track record when it comes to leaks than MLID, has only posted the block diagrams. Everything else is guesswork.

It is possible that those CU blocks he has used are, in reality, a pair of CUs or, another option, is that AMD is doubling the number of cores/shading units per CU, going from 64 to 128. They haven't increased that ratio in ages.

It doesn't help the other rumors, like AMD designing some parts to be able to be used for both desktop and laptop PCs (the bandwidth LPDDR5X provides doesn't even come close to GDDR6), or MLID's previous leaks, with quite different info (some of which he doesn't even believe).

And, on another note, how come 3 of the diagrams have 6 CUs per Shader Engine while the other one has only 5? Why that change? Is it a mistake from Kepler? After all, it doesn't make much sense to make that design change for one chip.

Anyways, it's too early to take anything from granted or make eny estimates. The only thing we can take from this is that it looks like AMD wants to compete in all sements once again, for better or worse.



Please excuse my bad English.

Former gaming PC: i5-4670k@stock (for now), 16Gb RAM 1600 MHz and a GTX 1070

Current gaming PC: R5-7600, 32GB RAM 6000MT/s (CL30) and a RX 9060XT 16GB

Steam / Live / NNID : jonxiquet    Add me if you want, but I'm a single player gamer.

JEMC said:
Jizz_Beard_thePirate said:

AMD Reportedly Preparing 16 GB Radeon RX 9070 GRE Graphics Card, Launch Expected In September-October

https://wccftech.com/amd-preps-16-gb-radeon-rx-9070-gre-launch-in-september-october/

I'm having a hard time figuring what AMD is trying to do with this one.

The 9070 GRE and the 9070 non-XT are already quite close in terms of performance, with the memory configuration being their biggest difference so, why change that to make it more complicated? At this point, it would make more sense to just kill the 9070GRE and cut the price of the base 9070... although AMD would need to finally be able to get enough of those chips to the cards could come close to MSRP.

The 9070GRE is a die-harvested part, cutting down the number of CU's by 8.

AMD must have enough "slightly" defective Radeon 9070's that they can justify pushing these into the market rather than throwing them into the bin... Or badging them as 9060XT's which erodes profit.

Darc Requiem said:

I'm not going to pretend to be hyper tech-savy but just on the face of it. The WCCF Tech article doesn't add up. A 12CU dGPU just make zero sense. That's what their mainstream current gen APUs top out at. Their weakest dGPU, the 9060 (non XT), has 28 CUs and that's been their low water mark since RDNA3 (7600S laptop GPU). I know there is a segment of the internet that has a huge hate boner for MLID. I get it, the guy can be smug at times, but he gets far too many things correct to be pulling things out of his rear like the average tech article commenter would have us believe. Given the alleged IPC improvements of RDNA5/UDNA a 24, not 12, CU GPU would have raster performance similar to a 9060 with substantially better RT and PT performance. That's believable.

We'll see the truth of the matter at some point. However the leaked GPU philosophy that AMD is seemingly going for is very smart. It allows them to use the same GPU dies across multiple product segments. The memory solution, especially with the low end SKUs, provides a potential good product for the entry level market. It would finally move us beyond the ridiculous 8GB frame buffer while being cheap enough from a business standpoint to actually come to market. That said, AMD often has good ideas on paper for their GPU business. For the sake of competition, I hope it pans out this time.


The guy is simply an idiot. I don't even care about how he comes across...
He will literally go back and delete the videos where his claims turned out to be false... And then in future videos constantly parades the fact he guessed one right.
https://www.reddit.com/r/BustedSilicon/comments/yo9l2i/collection_of_mlidmoores_law_is_dead_past_claims/

In-fact, Moores Law is Dead trends at about a 50% accuracy level... Which is a pretty poor statistic, especially when outlets like WCCFTECH quote him all the time and try to pass it off as fact.
https://www.reddit.com/r/Amd/comments/zvi9pg/mlid_accuracy_about_his_amd_leaks_a_70_claim/

A lot of his "guesses" are actually just "educated guesses" that anyone who has an idea about silicon and microprocessors can get right.

I mean... Prior to the 9th generation consoles launched or even being revealed... I made the "educated guess" that the next gen Xbox and Playstation consoles would come with 16GB of Ram... No more due to cost/performance benefits.
Many people thought we would get 32GB-128GB of Ram based on historical multiplying of DRAM with each successive console release.

It's simply an educated guess that I was absolutely bang-on with.

*****

As for the CU counts, we will need to wait for more information... UDNA is a fundamental architectural shift, AMD has already hinted it's looking to go the VLIW route again, which we saw the first with RDNA3 with the doubling of instructions per clock with each ALU. - So it may make sense to cut down the number of CU's to stay within transistor budgets and rely on the compiler for better throughput and utilization.

But, we will need to wait for more, RELIABLE information... And sadly, I refuse to accept "data" from Moores Law is Dead or Red Gaming Tech, both are not trustworthy based on historical trends and claims.




www.youtube.com/@Pemalite

Around the Network
Jizz_Beard_thePirate said:

Possible AMD RDNA 5 / UDNA GPU SKU Configs Point To 96, 40, 24, 12 CU Dies

https://wccftech.com/possible-amd-rdna-5-udna-gpu-sku-configs-point-to-96-40-24-12-cu-dies/

That would be a very odd split of dies in terms of CU or WGP count.

40, 24 and 12 are very close to each other, while on the other hand 96 is way above 40, that's too big of a gap, especially if those are supposed to be WGPs, not CU. 12CU also don't make sense unless that one would be APUs



JEMC said:

The biggest problem is that Kepler, that has a better track record when it comes to leaks than MLID, has only posted the block diagrams. Everything else is guesswork.

It is possible that those CU blocks he has used are, in reality, a pair of CUs or, another option, is that AMD is doubling the number of cores/shading units per CU, going from 64 to 128. They haven't increased that ratio in ages.

It doesn't help the other rumors, like AMD designing some parts to be able to be used for both desktop and laptop PCs (the bandwidth LPDDR5X provides doesn't even come close to GDDR6), or MLID's previous leaks, with quite different info (some of which he doesn't even believe).

And, on another note, how come 3 of the diagrams have 6 CUs per Shader Engine while the other one has only 5? Why that change? Is it a mistake from Kepler? After all, it doesn't make much sense to make that design change for one chip.

Anyways, it's too early to take anything from granted or make eny estimates. The only thing we can take from this is that it looks like AMD wants to compete in all sements once again, for better or worse.

I don't take MLID leaks as gospel but he clearly has inside source at AMD. Sony claimed his PS5 Pro video and Youtube took it down. Which I never understand why companies do that. Getting a video yanked just confirms the accuracy of the information. The problem I have with MLID is when he is wrong he refuses to take the L. IIRC I think it was the Intel Arrow Lake pricing where he was wrong and twisted himself into knots to avoid just saying "Hey I got this wrong." Things change in the tech space, some times you are going to get bad info or the company may change plans. 

Pemalite said:
JEMC said:

I'm having a hard time figuring what AMD is trying to do with this one.

The 9070 GRE and the 9070 non-XT are already quite close in terms of performance, with the memory configuration being their biggest difference so, why change that to make it more complicated? At this point, it would make more sense to just kill the 9070GRE and cut the price of the base 9070... although AMD would need to finally be able to get enough of those chips to the cards could come close to MSRP.

The 9070GRE is a die-harvested part, cutting down the number of CU's by 8.

AMD must have enough "slightly" defective Radeon 9070's that they can justify pushing these into the market rather than throwing them into the bin... Or badging them as 9060XT's which erodes profit.

Darc Requiem said:

I'm not going to pretend to be hyper tech-savy but just on the face of it. The WCCF Tech article doesn't add up. A 12CU dGPU just make zero sense. That's what their mainstream current gen APUs top out at. Their weakest dGPU, the 9060 (non XT), has 28 CUs and that's been their low water mark since RDNA3 (7600S laptop GPU). I know there is a segment of the internet that has a huge hate boner for MLID. I get it, the guy can be smug at times, but he gets far too many things correct to be pulling things out of his rear like the average tech article commenter would have us believe. Given the alleged IPC improvements of RDNA5/UDNA a 24, not 12, CU GPU would have raster performance similar to a 9060 with substantially better RT and PT performance. That's believable.

We'll see the truth of the matter at some point. However the leaked GPU philosophy that AMD is seemingly going for is very smart. It allows them to use the same GPU dies across multiple product segments. The memory solution, especially with the low end SKUs, provides a potential good product for the entry level market. It would finally move us beyond the ridiculous 8GB frame buffer while being cheap enough from a business standpoint to actually come to market. That said, AMD often has good ideas on paper for their GPU business. For the sake of competition, I hope it pans out this time.


The guy is simply an idiot. I don't even care about how he comes across...
He will literally go back and delete the videos where his claims turned out to be false... And then in future videos constantly parades the fact he guessed one right.
https://www.reddit.com/r/BustedSilicon/comments/yo9l2i/collection_of_mlidmoores_law_is_dead_past_claims/

In-fact, Moores Law is Dead trends at about a 50% accuracy level... Which is a pretty poor statistic, especially when outlets like WCCFTECH quote him all the time and try to pass it off as fact.
https://www.reddit.com/r/Amd/comments/zvi9pg/mlid_accuracy_about_his_amd_leaks_a_70_claim/

A lot of his "guesses" are actually just "educated guesses" that anyone who has an idea about silicon and microprocessors can get right.

I mean... Prior to the 9th generation consoles launched or even being revealed... I made the "educated guess" that the next gen Xbox and Playstation consoles would come with 16GB of Ram... No more due to cost/performance benefits.
Many people thought we would get 32GB-128GB of Ram based on historical multiplying of DRAM with each successive console release.

It's simply an educated guess that I was absolutely bang-on with.

*****

As for the CU counts, we will need to wait for more information... UDNA is a fundamental architectural shift, AMD has already hinted it's looking to go the VLIW route again, which we saw the first with RDNA3 with the doubling of instructions per clock with each ALU. - So it may make sense to cut down the number of CU's to stay within transistor budgets and rely on the compiler for better throughput and utilization.

But, we will need to wait for more, RELIABLE information... And sadly, I refuse to accept "data" from Moores Law is Dead or Red Gaming Tech, both are not trustworthy based on historical trends and claims.

Well as I mentioned above. I don't take his info as gospel but he gets enough right for me to take notice. Some of his things are to specific just to be educated guesses. As you stated though and I agree he just can't admit when he's wrong. Now Red Gaming Tech, I don't watch that guys videos at all. I can't remember him getting anything right. Moreover he'll put out a ten minute video that says nothing. Reminds of me of when politician gives a long answer to a question that never answers the question. A bunch of words, no substance. Truthfully the guy is in the wrong lane, he's jacked. He should put out fitness content. Unlike with his tech videos, he clearly knows what he is doing when it comes to hitting the gym. 



FADU announces PCIe Gen6 SSD controller enabling 28.5 GB/s speed and 9W power consumption

https://videocardz.com/newz/fadu-announces-pcie-gen6-ssd-controller-enabling-28-5-gb-s-speed-and-9w-power-consumption

AMD Wraith Prism and Spire CPU coolers discontinued for select Ryzen CPUs

https://videocardz.com/newz/amd-wraith-prism-and-spire-cpu-coolers-discontinued-for-select-ryzen-cpus

NVIDIA Confirms Rubin Chips Already In Fab & Ready For Volume Production In 2H 2026 While Gaming Posts Record $4.3 Billion Revenue In Q2

https://wccftech.com/nvidia-confirms-rubin-chips-already-in-fab-ready-for-volume-production-2h-2026-gaming-posts-record-4-3-billion-revenue/

Framework Laptop 16 will now let users upgrade graphics to RTX 5070

https://videocardz.com/newz/framework-laptop-16-will-now-let-users-upgrade-graphics-to-rtx-5070

Microsoft is promising to make Bluetooth audio much better in Windows 11

https://www.neowin.net/news/microsoft-is-promising-to-make-bluetooth-audio-much-better-in-windows-11/

Microsoft is promising to enable both microphone and high fidelity audio support using their currently supported Bluetooth LE Audio feature on a future Windows Update. 



                  

PC Specs: CPU: 7800X3D || GPU: Strix 4090 || RAM: 32GB DDR5 6000 || Main SSD: WD 2TB SN850

The Thursday gaming news, all in one long post:

SALES/PLAYER COUNTS & DEALS

This week, the Epic Store gives away these two games:

Next week, the free games will be The Battle of Polytopia and Monument Valley.

Steam has these deals:

There’s a new Humble Bundle: the Capcom Due of Justice: Ace Attorney and Mega Man Bundle, with up to 11 items to get (9 games and 2 coupons) during 26 days: https://www.humblebundle.com/games/duo-justice-ace-attorney-and-mega-man.

Fanatical has the new BYO Pedal to the Metal Bundle - Full Throttle, with 20 games to choose from for your 3+, 5+ or 7+ game bundles during 34 days: https://www.fanatical.com/en/pick-and-mix/build-your-own-pedal-to-the-metal-bundle.

SOFTWARE & DRIVERS

Epic's CEO Tim Sweeney wades in on the UE performance debate: 'The primary reason Unreal Engine 5-based games don't run smoothly on certain PCs or GPUs is the development process'
https://www.pcgamer.com/hardware/epics-ceo-tim-sweeney-wades-in-on-the-ue-performance-debate-the-primary-reason-unreal-engine-5-based-games-dont-run-smoothly-on-certain-pcs-or-gpus-is-the-development-process/
In the world of PCs and PC gaming, debates are usually quite evenly distributed. AMD vs Intel. AMD vs Nvidia. WASD vs arrow keys. Cats vs dogs. You'll find arguments for either side most of the time. However, when it comes to Unreal Engine 5, comments from PC gamers are mostly on the side of "it just runs really badly." Epic Games has tried to counter such claims many times before, but with the debate still running, CEO Tim Sweeney has stepped in to say that the main reason why UE5 games don't run well is how they're developed in the first place.
He said this at the recent Unreal Fest event in Seoul, as reported by Korean site This is Game (via RedGamingTech). Press members asked about what Epic Games was doing to address optimisation issues for games developed on UE5. "The primary reason Unreal Engine 5-based games don't run smoothly on certain PCs or GPUs is the development process," replied Sweeney (alas, my grasp of Korean is non-existent, so hopefully the machine translation has got it spot on).
More specifically, he suggested that one particular problem is the choice of PC platform used in the formative stages of a game's creation. "Many developers begin by developing games for high-end hardware, then optimize and test on lower-spec devices in the final stages."
>> That doesn’t explain why users with the highest-end systems available also have those problems, or why they happen with console ports.

MODS, EMULATORS & FAN PROJECTS

In the wake of shutting down a highly popular mod, Final Fantasy 14 director Yoshi-P shares deeply earnest blog about 'mods, their use, and the culture surrounding them'
https://www.pcgamer.com/games/final-fantasy/in-the-wake-of-shutting-down-a-highly-popular-mod-final-fantasy-14-director-yoshi-p-shares-deeply-earnest-blog-about-mods-their-use-and-the-culture-surrounding-them/
In case you're unfamiliar with the Final Fantasy 14 community, recently, a mod named Mare Lamentorium—which had many thousands of users at its peak—was sent a cease & desist order. The mod in question allowed players to share their modded appearances with each other, either via an individual code, or by allowing players to join 'syncshells'.
From a community standpoint, Mare was used in two ways—either for innocuous roleplayers or outfit enthusiasts to customise their character, say, by giving a human model pointy ears—or, and I say this without judgement, smut. There is an enormous underground modding scene in Final Fantasy 14, covering both adult and non-adult content.
The biggest issue being that addons are against the terms of service of Final Fantasy 14—however, Square Enix will only take action if a player is obviously cheating, or if they are reported for mod use. In other words, players don't get banned for having a DPS meter, they get banned for using that DPS meter to chew out their party members in-game.
Community uproar around Mare Lamentorium's been loud, however—loud enough that game director Naoki Yoshida (Yoshi-P) has felt the need to create an extensive blog post on the subject.
At the start of this blog post, Yoshi-P asks media outlets (hey, that's me!) to "Please do not summarize or provide only excerpts." This obviously puts me in a tough spot, given it's my job to add context to videogame news—especially as someone with thousands of hours in this game, and a foot in its roleplay community.
>> While the article continues and does exactly what Yoshi-P asked them not to, I’ll leave it there.

GAMING NEWS

Riot leaked its next 2XKO fighter in the silliest way possible before throwing its hands up and saying 'oopsy daisy' with an announcement post
https://www.pcgamer.com/games/fighting/riot-leaked-its-next-2xko-fighter-in-the-silliest-way-possible-before-throwing-its-hands-up-and-saying-oopsy-daisy-with-an-announcement-post/
Leaks are part and parcel of a videogame's development at this point. Do I agree with that? Of course not, but you'd be hard pressed to find a big game that hasn't sprung a little leak here and there at some point. Sometimes it's dataminers getting their hands on a beta, or in more unfortunate cases (like GTA 6), it's a big ol' hack job.
Sometimes, though, the call comes from inside the house. That's exactly what happened in the case of 2XKO—Riot's upcoming League of Legends tag fighter—when an unannounced character showed up on the developer's goddamn support page of all places.

Hinako Shimizu's arc in Silent Hill f is shaped by the Japanese women's rights movement of the 1960s and depicts 'how she musters the courage to combat' repression she encounters
https://www.pcgamer.com/games/horror/hinako-shimizus-arc-in-silent-hill-f-is-shaped-by-the-japanese-womens-rights-movement-of-the-1960s-and-depicts-how-she-musters-the-courage-to-combat-repression-she-encounters/
Hinako Shimizu is shaping up to be one of the more intriguing protagonists I've encountered in a Silent Hill game. Even after three hours of playing through the beginning of Silent Hill f I still can't quite figure out whether she's as innocent and selfless as she appears. But what is clear is that, like many characters present in Silent Hill games, she's been through a hell of a lot of trauma.

Hideo Kojima questionably visits Saudi Arabia to drop the news that he knows just what Death Stranding 3 should be but he wants someone else to make it
https://www.pcgamer.com/games/action/hideo-kojima-questionably-visits-saudi-arabia-to-drop-the-news-that-he-knows-just-what-death-stranding-3-should-be-but-he-wants-someone-else-to-make-it/
Great news: Hideo Kojima already knows just what he wants Death Stranding 3 to be. Bad news: some other poor sod is gonna have to make it, because Kojima's very busy converting Willem Dafoe into an Ogg Vorbis file or whatever it is he does these days.

Here's every game that Silksong's surprise release has delayed as indies scramble to escape its powerful aura
https://www.pcgamer.com/games/heres-every-game-that-silksongs-surprise-release-has-delayed-as-indies-scramble-to-escape-its-powerful-aura/
"From the town of Lincoln, Nebraska, with a sawed-off .410 on my lap," Bruce Springsteen crooned in his magnificent 1982 ballad Nebraska. "Through to the badlands of Wyoming, I killed everything in my path." The Boss was of course relating the tale of infamous mass murderer Charles Starkweather, but he might as well have been talking about Hollow Knight: Silksong, the upcoming metroidvania that has now claimed at least eight victims since its surprise release date announcement earlier this month.
Nobody's actually getting murdered here, to be clear: When I say "victims" I mean "other videogames," and by "claimed" I mean that Silksong's looming launch has led those games to adjust their own release plans so as not to be caught in the blast radius. It's been quite a show, and while this sort of release date scramble isn't unprecedented—Starfield made something of a mini-splash a couple years ago—I haven't seen this kind of fracas since Cyberpunk 2077 started throwing elbows back in 2020.

Blizzard promises to add 'speed, power, and fun' back to WoW Legion Remix after disappointing players with a time-gated grind
https://www.pcgamer.com/games/world-of-warcraft/blizzard-promises-to-add-speed-power-and-fun-back-to-wow-legion-remix-after-disappointing-players-with-a-time-gated-grind/
No one is safe from the ire of players who feel like their toys were ripped away from them. Not even Blizzard when it tried to balance World of Warcraft's next Remix event after players overwhelmingly loved how unbalanced the last one was.

Judas must be getting closer, because Ken Levine just rolled out its first-ever dev log and some new art
https://www.pcgamer.com/games/fps/judas-must-be-getting-closer-because-ken-levine-just-rolled-out-its-first-ever-dev-log-and-some-new-art/
Three years after Shock recidivist Ken Levine formally unveiled his next game as Judas, we still don't know much about it, except that it sure looks like a new BioShock. But that situation will hopefully be changing soon: Levine posted the very first Judas dev log today, and said he hopes "to communicate more frequently to update you with new details of what we are working on."

Clair Obscur: Expedition 33 director adds another DLC teaser to the pile: 'We may be cooking'
https://www.pcgamer.com/games/rpg/clair-obscur-expedition-33-director-adds-another-dlc-teaser-to-the-pile-we-may-be-cooking/
I don't blame any of the developers behind Clair Obscur: Expedition 33 for not immediately thinking about DLC plans just as their game exploded in popularity and hit sales numbers in the millions earlier this year. But in the last few months they've acknowledged the hunger for DLC in every way but a direct confirmation and at this point I'd just like to know if those plans are real.
(...)
That brings us to a comment from director Guillaume Broche this week in an interview with YouTuber MrMattyPlays. When asked about plans to release DLC, Broche said, "There may be," and that it's "a bit too early to say."

Final Fantasy 14's incoming choose-your-own difficulty will be 'the general direction' going forward, but Yoshi-P suggests they might need to trim future patches to compensate
https://www.pcgamer.com/games/final-fantasy/final-fantasy-14s-incoming-choose-your-own-difficulty-will-be-the-general-direction-going-forward-but-yoshi-p-suggests-they-might-need-to-trim-future-patches-to-compensate/
Final Fantasy 14's gonna be shaking up how it handles difficulty come patch 7.35, after a mini-disaster with patch 7.25's Occult Crescent back in May. To put a long story short, Creative Studio 3's somewhat fumbled approach to content difficulty came to a head; the Forked Tower somehow managed to thread the needle, becoming a raid made for no-one.
(...)
That's changing with 7.35's Pilgrim's Traverse, which'll have all sorts of variable difficulty content—you'll be able to tackle all 100 floors of the deep dungeon with a matchmade group, but you'll then be able to challenge a "quantum" difficulty boss via offerings. Give more offerings, get a harder butt-kicking.
>> He also said that we won’t have to wait long for the next expansion.

Kevin Spacey got 'a little weird' on the set of Call of Duty: Advanced Warfare: 'We all had to fake laugh'
https://www.pcgamer.com/games/call-of-duty/kevin-spacey-got-a-little-weird-on-the-set-of-call-of-duty-advanced-warfare-we-all-had-to-fake-laugh/
2014's Call of Duty: Advanced Warfare will always be the game that made us "Press F to pay respects", a weird moment of interactivity that undercut what was meant to be an emotional scene. Even when we're all gone, the meme will live on.
If there's one other thing that you'll probably remember about Sledgehammer's foray into future warfare, it's that it starred Kevin Spacey. (...)

The MMO granddaddy you've never heard of is celebrating its 35th anniversary, and it still pulls in 'a few thousand active players' and needs a staff of 40 to run: 'We want to be able to survive for another 35 years'
https://www.pcgamer.com/games/mmo/the-mmo-granddaddy-youve-never-heard-of-is-celebrating-its-35th-anniversary-and-it-still-pulls-in-a-few-thousand-active-players-and-needs-a-staff-of-40-to-run-we-want-to-be-able-to-survive-for-another-35-years/
If any one thing can claim responsibility for sending me down the dark, dark path of writing for a living, it's probably MUDs—text-based MMOs from an era before 3D graphics that somehow still endure today. It wasn't school, it wasn't uni, it wasn't a mentor's encouragement or some precocious childhood proclivity. I just spent most of my adolescence typing a lot and it just kind of kept happening as I approached adulthood.
And one of the granddaddies of MUDs as a genre is GemStone IV, currently marking its 35th anniversary. It's bigger than you might think: after starting "on the early ISPs—Genie, Prodigy, AOL" all the way back in 1990 and finally moving to the web proper in 1997, it still manages to pull in "a few thousand active players" in the modern era, per product manager Wyrom.

Bethesda remembers it made Starfield, promises new 'space gameplay' in back half of random video that doesn't even mention Starfield in the title
https://www.pcgamer.com/games/rpg/bethesda-remembers-it-made-starfield-promises-new-space-gameplay-in-back-half-of-random-video-that-doesnt-even-mention-starfield-in-the-title/
Oh, Starfield. Since the release of—and lukewarm response to—the Shattered Space expansion almost a full year ago, Bethesda's been very quiet about what it's got planned for the game that was once anticipated enough to send Baldur's Gate 3 scurrying for cover. In fairness, the studio's had other irons in the fire, most notably the much-better-received Oblivion remaster that hit earlier this year.
(...)
Well take heart, true believers, because someone has remembered the game exists. In a recent chat with dev Tim Lamb on Bethesda's own YouTube channel, it was confirmed that, yep, there is something cooking for Starfield in the relatively near future. "I'm really excited for players to see what the team's been working on," said Lamb. "We have some cool stuff coming, including free updates and features that players have been asking for, as well as a new DLC story."
>> Kind of harsh headline given that this cool stuff may be linked to the interstellar travel PCGamer posted about earlier this month.

'We have become experts at putting Poland into games': Cronos: The New Dawn is set in a district of Kraków simply because Bloober wanted to 'show our crib to the audience'
https://www.pcgamer.com/games/horror/we-have-become-experts-at-putting-poland-into-games-cronos-the-new-dawn-is-set-in-a-district-of-krakow-simply-because-bloober-wanted-to-show-our-crib-to-the-audience/
Cronos: The New Dawn takes place in a post-apocalyptic version of the Kraków district of Nowa Huta during the 1980s. It seems like an oddly specific location, that is, until you actually ask the devs over at Bloober Team why they decided to set their upcoming survival horror game here.



Please excuse my bad English.

Former gaming PC: i5-4670k@stock (for now), 16Gb RAM 1600 MHz and a GTX 1070

Current gaming PC: R5-7600, 32GB RAM 6000MT/s (CL30) and a RX 9060XT 16GB

Steam / Live / NNID : jonxiquet    Add me if you want, but I'm a single player gamer.

Jizz_Beard_thePirate said:

Damn! Even with their gaming GPUs doing better than ever, the proportion is insane.



Please excuse my bad English.

Former gaming PC: i5-4670k@stock (for now), 16Gb RAM 1600 MHz and a GTX 1070

Current gaming PC: R5-7600, 32GB RAM 6000MT/s (CL30) and a RX 9060XT 16GB

Steam / Live / NNID : jonxiquet    Add me if you want, but I'm a single player gamer.