The 2 gig wall from 32 bit games & Civ 6

Arioch, I think you're over-estimating the number of people still on 32 bit.

The Steam survey isn't perfect, but it's a solid representation of what gamers use. It has just over 10% of their gamers on 32-bit systems.
http://store.steampowered.com/hwsurvey

When you think about it, if Firaxis has potentially 2 systems to support (32 and 64 bit), then they need to divide that support in half... it's not worth dividing the support by that much to cater to a 10% audience. That's too costly for not enough return.

Many games are now going the 64-bit exe only now. So it would not be surprising to me at all to see Civ VI abandon 32-bit. Unless they want the Chinese market.
 
Arioch, I think you're over-estimating the number of people still on 32 bit.

The Steam survey isn't perfect, but it's a solid representation of what gamers use. It has just over 10% of their gamers on 32-bit systems.
http://store.steampowered.com/hwsurvey

When you think about it, if Firaxis has potentially 2 systems to support (32 and 64 bit), then they need to divide that support in half... it's not worth dividing the support by that much to cater to a 10% audience. That's too costly for not enough return.

Many games are now going the 64-bit exe only now. So it would not be surprising to me at all to see Civ VI abandon 32-bit. Unless they want the Chinese market.

Right and 2K is looking at a sales / shelf life between 2016 -2022 when even fewer people will be on 32-bit and current 32-bit users upgrade.
 
Big J Money, nailed it. Stardock looked at the data and went with 64 bit only version of Gal Civ III. If I was a developer I would not cut my develop team assets in half to cater to 10% of the market. It makes more sense to have a solid foundation in 64 bit knowing that 32 bit OS will go the way of defunct operating systems eventually.
 
Doing a 32-bit and a 64-bit version is not cutting team assets in half by any stretch of the imagination. For the most part, you're literally just flipping a switch in the compiler. It's not a separate code base.
 
I hope its 64bit only but damn i would like it if it had Vulkan support. I really don't want to downgrade to windows 10 just for DX12... We need an open API to takeover badly.
 
Doing a 32-bit and a 64-bit version is not cutting team assets in half by any stretch of the imagination. For the most part, you're literally just flipping a switch in the compiler. It's not a separate code base.

Games are written in languages such as C / C++ in which the sizes of some of the primitives are subject to change between 32 bit & 64 bit.
Pointer sizes definitely change between 32 bit & 64 bit.
And with any libraries it really needs to be a clean vertical slice for compatibility.

It's not like Java / C# which abstracts away pointers and has defined primitives to be the same sizes regardless of what bit OS it's running on.
 
10% audience for Civ is about a million users. Since selling additional copy of digital product don't require additional resources, losing these people is pure money loss. Let's take into account discounts, regional pricing, share of people who will upgrade their hardware, etc. it's still about $15M of pure loss. Give me $10M and I'll optimize the game memory consumption to hell, it will work in 32-bit like a charm.

But I still could imagine the situations where 64-bit could be worth it. In Firaxis plans using the engine in future games, for example, etc. We don't know for sure. Also, this could be just emotional decision - game developers may not like to play with memory optimizations, they want different things to work on.
 
10% audience for Civ is about a million users. Since selling additional copy of digital product don't require additional resources, losing these people is pure money loss. Let's take into account discounts, regional pricing, share of people who will upgrade their hardware, etc. it's still about $15M of pure loss. Give me $10M and I'll optimize the game memory consumption to hell, it will work in 32-bit like a charm.

But I still could imagine the situations where 64-bit could be worth it. In Firaxis plans using the engine in future games, for example, etc. We don't know for sure. Also, this could be just emotional decision - game developers may not like to play with memory optimizations, they want different things to work on.

No, 2k is going to be looking at what the 10% will be like by 2022, when Civ7 hits.
And it likely won't be 10%

The first wave of people who buy this game will be hardcore. Many people are already talking about upgrading just in-time for the game. The people still on 32-bit probably aren't motivated enough to upgrade because there's no games they play that require anything more, and they will likely pick up Civ6 on a steam summer sale in the summer of 2019 when Civ6 - Ed's Wrath is released. By that time, they may be 64 bit already just from network effects of more and more games going that way.

But the first 1-3 million in Civ6 sales in its Q1 on the market will be to hardcore fans and those people will be ready for it.

This computer I'm using is on a gen1 i7 on an older socket so I can't even upgrade to a new i7, but I got it OCd and running with a GTX970 it runs Civ5 and most modern games great. I'll run Civ6 on this oldie for a while as well. The only change I made 2 years ago was get Win7 64-bit from my old 32-bit Vista so I could up my RAM from 3GB and increase performance for my general gaming at the time.
 
You can't indefinetely await for everyone to catch up to the latest technology. The last comers to 64bit will do it once they are forced to do so to play what they want to play.

That has always been this way in gaming.

If the technology allows improvements, the gain in popularity due to the product quality can outweigh the cost of having a minority not able to play. That is how some game companies works when they always make game requiring top hardware. Their sale point in that case are the graphics/technology.

Technology sells. What would be bad is using 64 bit with no gain.
 
No, 2k is going to be looking at what the 10% will be like by 2022, when Civ7 hits.

People who buy games several years after release have too huge discounts to be counted. Also note - Civ6 has strategic view mode, which is likely designed to work on older hardware and tablets. Many of them have 32-bit OS.

The first wave of people who buy this game will be hardcore. Many people are already talking about upgrading just in-time for the game. The people still on 32-bit probably aren't motivated enough to upgrade because there's no games they play that require anything more, and they will likely pick up Civ6 on a steam summer sale in the summer of 2019 when Civ6 - Ed's Wrath is released. By that time, they may be 64 bit already just from network effects of more and more games going that way.

"Hardcore" civ players and hardcore gamers in general are different audiences. Strategic games have lower system requirements and so we don't have such cutting-edge computers.

Most people just need to pay some money to get a new 64-bit OS as their CPU are already 64-bit, or get the Win10 64-bit free, but without the smooth transition from Win7/8 32bit to Win10 32-bit.

If you have 32-bit Win7/Vista, it upgrades to Win10 32-bit too, unfortunately. So, it doesn't drop the rate of 32-bit OSes. And since currently most of OS sales are OEM, it's unlikely people will buy new OS without buying new hardware.
 
@Stealth_nsk can't people choose to upgrade to Win10 64-bit? They just loose the smooth transition.

Was the same for me when I upgraded from Vista 32 to Win 7 64

Studied the question now. By default offer updates 32-bit Windows 7/8 to 32-bit Windows 10. It seems possible to reinstall with 64-bits, but it's the same as with previous Windows versions, so it's unlikely people with 32-bit OSes will do so. Most of them are likely to have OEM Windows and don't want to screw it.
 
Truth to be told, those with 32bit systems probably don't have DX11 video cards either, and I really doubt that Civ6 will support DX9, as it did 5 years ago.

With decline of WinXP, both DX9 and 32bit OS are dying out, as platforms for games.
 
Games are written in languages such as C / C++ in which the sizes of some of the primitives are subject to change between 32 bit & 64 bit.
Pointer sizes definitely change between 32 bit & 64 bit.
And with any libraries it really needs to be a clean vertical slice for compatibility.

It's not like Java / C# which abstracts away pointers and has defined primitives to be the same sizes regardless of what bit OS it's running on.

Yes; that's true. I'm familiar with how C/C++ handles data. However, my point remains. If they're building for both 32-bit and 64-bit there's still little more to it than selecting an option from a drop-down list.

Any dependencies they're not compiling themselves they'd already have 32 and 64 bit versions of. The only primitive that really changes size between 32 and 64 bit in MSVC (besides pointers, which naturally become 8 bytes in size in order to properly address all possible memory) is long, which goes from 4 bytes (same as int, because of annoying historical reasons) to a full 8 bytes. And if the size of an integer matters, one should really be using the stdint.h types anyway, not the built-in types which aren't rigidly defined in the C++ standard and can change from compiler to compiler (especially char, which is all over the place).

It's a minuscule amount of extra effort for a studio Firaxis's size, especially one where the compiling and bundling process is going to be mostly automated anyway.
 
Doing a 32-bit and a 64-bit version is not cutting team assets in half by any stretch of the imagination. For the most part, you're literally just flipping a switch in the compiler. It's not a separate code base.

This is not a indie, the point is that the game should scale equally, and performance of 32 bit must be nearly as good as 64 at least in non huge maps.

Ofc making the executable 64 or 32 is easy, but if the core structure is 64 bit, there will be problems, same if the code is made for 32, the 64 version made with "one click" will just be not optimized for 64 so basically not that different.
 
Yes; that's true. I'm familiar with how C/C++ handles data. However, my point remains. If they're building for both 32-bit and 64-bit there's still little more to it than selecting an option from a drop-down list.

Any dependencies they're not compiling themselves they'd already have 32 and 64 bit versions of. The only primitive that really changes size between 32 and 64 bit in MSVC (besides pointers, which naturally become 8 bytes in size in order to properly address all possible memory) is long, which goes from 4 bytes (same as int, because of annoying historical reasons) to a full 8 bytes. And if the size of an integer matters, one should really be using the stdint.h types anyway, not the built-in types which aren't rigidly defined in the C++ standard and can change from compiler to compiler (especially char, which is all over the place).

It's a minuscule amount of extra effort for a studio Firaxis's size, especially one where the compiling and bundling process is going to be mostly automated anyway.
I often wonder why people make assumptions about the complexity of a codebase and / or dependencies - especially when they're technical themselves.

I think it's safe to say that while some of the basic principles might seem straightforward, working that out across a codebase the size of your average game engine is perhaps not that straightforward.
 
Moderator Action: Several posts related to 32 bit vs 64 bit merged into this thread.
Thread reopened.
 
Back
Top Bottom