The Citrix Session: Unlocking New Possibilities with Intel GPUs in Citrix Environments

In today’s dynamic work landscape, remote access and virtual desktop solutions have become essential for organizations striving for flexibility and productivity. Citrix has been at the forefront of empowering businesses with innovative technologies, and now, with Intel’s new roadmap of GPU offerings, users can expect even more versatility and performance in their virtual desktop experiences.

Host: Andy Whiteside
Co-host: Bill Sutton
Co-host: Todd Smith

Transcript

WEBVTT

1
00:00:02.520 --> 00:00:12.089
Andy Whiteside: Everyone welcome to episode 1 59 of the Citrix session. I'm your host, Andy Whiteside. That's right, guys. A 159 of these so far no end in sight.

2
00:00:12.290 --> 00:00:21.327
Andy Whiteside: I I'll just take this opportunity to thank Todd and Jeremy and Monica for helping Bill. And I do this. This is something special. We do.

3
00:00:21.906 --> 00:00:45.979
Andy Whiteside: I don't think we're getting as many folks listening as we want, but that there are hundreds and hundreds and hundreds of people listening, which is good and we're really excited to be adding, you know, value the way we do. I've got Bill Sutton with the bill has a newer title. I don't know if we talked about this yet or not, but Bill is now the director of our modern workspace practice our business units which means now, Bill is responsible for all things. Citrix, including

4
00:00:46.090 --> 00:00:50.739
Andy Whiteside: endpoints and services and other related technology bill. How's it going.

5
00:00:51.272 --> 00:00:52.679
Bill Sutton: It's going. Well, Andy.

6
00:00:53.200 --> 00:00:56.680
Andy Whiteside: I like. I like the new background. I just found out we had a new background just the other day.

7
00:00:56.680 --> 00:00:59.029
Bill Sutton: Yeah, I thought I tried out for this, so we'll see.

8
00:00:59.030 --> 00:00:59.580
Andy Whiteside: Good

9
00:01:00.317 --> 00:01:09.710
Andy Whiteside: with us. We also have Todd Smith. Todd's been a loyal friend of Zintte. He gets it, man. He just gets that Zintegra is here to to add value in the Citrix space. Todd, how are you.

10
00:01:09.960 --> 00:01:11.420
Todd Smith: I'm doing well. How are you?

11
00:01:11.630 --> 00:01:13.810
Andy Whiteside: Good are you in Canada, or you at home?

12
00:01:14.450 --> 00:01:19.400
Todd Smith: I'm home this week last week I was up in Ottawa, and next week I'm out in Vancouver.

13
00:01:19.400 --> 00:01:20.020
Andy Whiteside: Hmm!

14
00:01:20.640 --> 00:01:23.099
Andy Whiteside: Are you getting like a like a lot of flight miles.

15
00:01:24.440 --> 00:01:26.165
Todd Smith: Yep, yep, it's

16
00:01:26.840 --> 00:01:40.771
Todd Smith: It's been interesting. Getting from Boston up to ottawa was, you know, less than an hour flight type of thing it's on a really. There's a kind of a hidden gem airline covering Canada called porter airlines. It's

17
00:01:41.840 --> 00:01:43.560
Todd Smith: really interesting. It's a

18
00:01:43.670 --> 00:01:49.390
Todd Smith: they fly. A lot of dash eights. So the turbo props, which is always a fun adventure.

19
00:01:49.390 --> 00:01:49.830
Andy Whiteside: I.

20
00:01:50.160 --> 00:01:53.837
Andy Whiteside: Imagine your big old body and a turbo prop that makes you scared.

21
00:01:54.120 --> 00:01:57.579
Todd Smith: Little bit of a challenge, but you know it's it's a short flight.

22
00:01:57.580 --> 00:01:59.810
Andy Whiteside: Do they make you sit in a certain part of the plane.

23
00:02:00.240 --> 00:02:01.920
Todd Smith: No, no.

24
00:02:02.580 --> 00:02:08.020
Todd Smith: no! That that that was the old days where they used to have to like balance everything out and all that stuff. But.

25
00:02:08.020 --> 00:02:08.580
Andy Whiteside: Coming!

26
00:02:08.710 --> 00:02:15.960
Todd Smith: The newer ones are are not bad, and the thing is, the thing about the dash hates, though, is that they can turn off both engines and still collide

27
00:02:16.900 --> 00:02:21.870
Todd Smith: down to the to the ground effectively. So that's always nice to know.

28
00:02:21.870 --> 00:02:23.600
Andy Whiteside: Well, I hope I never have to find that out.

29
00:02:23.600 --> 00:02:27.627
Bill Sutton: I would think that would be a little unsettling, but it's good to know that they can do that.

30
00:02:27.840 --> 00:02:29.089
Todd Smith: Well, jet, scope.

31
00:02:29.730 --> 00:02:30.930
Bill Sutton: Yeah, I know, I know.

32
00:02:31.330 --> 00:02:31.870
Bill Sutton: Yeah. But.

33
00:02:31.870 --> 00:02:33.580
Andy Whiteside: Because of the weight of the jets.

34
00:02:34.410 --> 00:02:38.660
Todd Smith: Good part of it's the the thing that keeps the jets in the air is the thrust.

35
00:02:38.660 --> 00:02:39.580
Bill Sutton: Yeah. Hmm.

36
00:02:40.910 --> 00:02:42.800
Andy Whiteside: Well, guys, we we picked a couple of

37
00:02:43.230 --> 00:02:49.509
Andy Whiteside: blogs here a minute ago to talk through, and I have lost them on my screen. Let me bring it back up here. We're going to talk about the intel graphics

38
00:02:49.680 --> 00:03:04.957
Andy Whiteside: blog. And I'm gonna put Todd on the spot right right fast here right quick. The name of the the name of the blog, and it's from 6 days ago is unlocking new possibilities with intel gpus in citrix environments. I don't think I've talked about this in forever, in terms of

39
00:03:05.740 --> 00:03:27.270
Andy Whiteside: a podcast but I talk about all time with clients, and I'll I'll set the conversation up by saying this. You know every desktop, every PC. Every laptop that you've owned in the last 20 years. Ish has had a Gpu in it. What makes people think they're gonna put this into a virtual world and not give it some aspect of a Gpu and rely on CPU only for the processing and the images is absolutely insane.

40
00:03:27.542 --> 00:03:44.969
Andy Whiteside: I get with customers all the time. They immediately jump into like autocad, and all these 3D rendering stuff, and I'm like that's great. But if you just want snappy experiences in office apps and other apps, including browsers, you need some form of Gpu involved. Hey, Todd, before I get too too far into this. The author of this one is

41
00:03:45.290 --> 00:03:46.080
Andy Whiteside: Poop.

42
00:03:47.780 --> 00:03:50.310
Todd Smith: I believe it's L. Joe.

43
00:03:50.792 --> 00:03:58.490
Todd Smith: Ljob Ben Gulik. He's one of our product managers focus on obviously the graphic side.

44
00:03:58.839 --> 00:04:13.760
Todd Smith: And you know it. It's been interesting that that this topic keeps coming up because I'd almost wanna say that they they need the industry as a whole needs to eliminate any graphics processing unit because you think of it's a gaming or a

45
00:04:13.900 --> 00:04:17.350
Todd Smith: architecture autocat type. Only solution.

46
00:04:17.720 --> 00:04:25.240
Todd Smith: The biggest use case for Gpus is number crunching. Yeah, that's what it does. And it does number crunching in the

47
00:04:25.713 --> 00:04:35.430
Todd Smith: the non integer numbers. Right? So it yes, it does do an awful lot associated with graphics because graphics is typically a non integer based

48
00:04:35.660 --> 00:04:36.880
Todd Smith: format

49
00:04:37.140 --> 00:04:38.879
Todd Smith: for processing. But

50
00:04:39.230 --> 00:04:43.880
Todd Smith: the biggest thing that they can do is they can push a ton of

51
00:04:43.940 --> 00:04:46.360
Todd Smith: calculations through a Gpu

52
00:04:46.783 --> 00:04:50.506
Todd Smith: that you and it offloads off of the CPU.

53
00:04:50.920 --> 00:05:04.689
Andy Whiteside: Everything the CPU should not waste its time doing. The gpus can do. And that just makes for a we're talking about turbo prop planes went. It's like a turbo charger. It's like having a, you know, 4 cylinder, 6 cylinder car, 8 cylinder car with a turbo charger.

54
00:05:05.040 --> 00:05:11.659
Andy Whiteside: That thing makes a massive difference when it's needed. And in today's workloads it's needed a lot whether you realize it or not.

55
00:05:12.980 --> 00:05:21.549
Andy Whiteside: Absolutely. Bill, so general thoughts on this Gpu conversation based on what I said, what Todd said, where it comes up and where it comes up. Not enough.

56
00:05:22.100 --> 00:05:42.656
Bill Sutton: Yeah, it doesn't come up enough for sure. And I think a lot of that has had to do with costs and somewhat with cost and somewhat with the a lot with just lack of understanding. I mean, today. Like to your point, Andy, you buy a laptop just a generic off the shelf laptop. It's gonna have some form of a Gpu, whether it's part of the same chip that's running the CPU or not.

57
00:05:42.920 --> 00:06:06.550
Bill Sutton: It's got some graphics cap some graphics processing capabilities and met number crutching, to use, to use Todd's words, capabilities within it, and even things as simple and seemingly benign as as a Powerpoint presentation, and the transitions you put in those are leveraging Gpu like functionality. So if you're if you're just publishing office or you're running office within a Vdi workload.

58
00:06:06.610 --> 00:06:12.379
Bill Sutton: and you don't have a Gpu office is probably not gonna perform quite as well as it would with the Gpu.

59
00:06:12.380 --> 00:06:37.190
Andy Whiteside: It's not going to, I mean, forget about the fact you got latency across the delivery mechanism of however much latencies across that wire that you're connecting from. It's not snappy. It's not as snappy on the back end, it's, you know, going back to the turbo idea. It's like, if we fast forward. 10 years from now every car is gonna be a smaller form factor engine with a turbo charger. And then all of a sudden you show up and just jump into a 4 cylinder Ford Ranger from 1989. You're like, this thing's a dog.

60
00:06:37.980 --> 00:06:51.250
Bill Sutton: Yes, like when I went from my 6 to cylinder standard sedan to a 4 cylinder sedan, I had to press on the gas a whole lot harder to get it to go as fast or as accelerate as quickly. There's a really, you know, a major difference.

61
00:06:51.390 --> 00:06:55.800
Andy Whiteside: Yeah, hey, Todd, do you wanna talk about the relationship between Citrix and Intel? During all this time.

62
00:06:56.050 --> 00:07:07.130
Todd Smith: Yeah. So so Citrix and Intel have had a longstanding relationship. Obviously, chip manufacturers have a have a huge influence intel was also one of the first that

63
00:07:07.740 --> 00:07:17.019
Todd Smith: kind of started putting virtualization into the silicon with the Vp. Pro chipsets. And they've also done an awful lot with the graphics, but they're

64
00:07:17.260 --> 00:07:20.319
Todd Smith: but oftentimes they're overshadowed by some of the other

65
00:07:20.763 --> 00:07:30.566
Todd Smith: companies out there that focus just on graphics. Right? So think of it. Back in the days of silicon graphics. And even in modern times, within video.

66
00:07:30.980 --> 00:07:38.050
Todd Smith: you know, Intel has a ton of really smart engineers that are that are building out solutions

67
00:07:38.140 --> 00:07:49.049
Todd Smith: that are going and interacting directly with the CPU right? So it it made a lot of sense to to partner up with Intel and keep and keep maintaining that relationship with Intel

68
00:07:49.816 --> 00:07:52.220
Todd Smith: and and be able to continue to

69
00:07:52.290 --> 00:07:59.029
Todd Smith: to, to co-develop and work on projects together, work on bringing new features to the world.

70
00:08:00.010 --> 00:08:20.614
Andy Whiteside: That makes total sense. As we move from endpoint computing to delivered computing. So deployed to delivered computing whether it's like we're gonna talk about here in the very beginning, a physical computer delivering across a wire to another physical entity of some type, physical physical device or whether it's a virtualized instance running on top of some hypervisor layer or in the cloud potentially.

71
00:08:21.740 --> 00:08:22.410
Andy Whiteside: go ahead. Sorry.

72
00:08:22.410 --> 00:08:29.719
Todd Smith: And I think, Andy, it also drives towards the the intelligence built into the Citrix stack to to

73
00:08:29.900 --> 00:08:39.980
Todd Smith: leverage a graphics card that's on a local endpoint on the remote endpoint as much as possible by offloading, by doing graphics, offloads and things like that.

74
00:08:40.330 --> 00:08:55.290
Todd Smith: But yet, when the user goes over to more of a thin client experience that doesn't have the High End graphics cards on the endpoints and really kind of adjusting that on the fly and be able to say, you know what you're still gonna get the same experience.

75
00:08:55.630 --> 00:09:01.110
Todd Smith: regardless of what endpoint you're on, and we're going to be smart enough in the back end, to

76
00:09:01.430 --> 00:09:05.070
Todd Smith: to put those resources to use wherever they happen to be.

77
00:09:05.070 --> 00:09:05.550
Andy Whiteside: Yeah, but.

78
00:09:05.550 --> 00:09:06.060
Todd Smith: We're.

79
00:09:06.060 --> 00:09:08.890
Andy Whiteside: Possible experience the most efficient way.

80
00:09:08.940 --> 00:09:12.309
Andy Whiteside: And and in this world, potentially the safest way.

81
00:09:12.610 --> 00:09:13.380
Todd Smith: Absolutely.

82
00:09:14.126 --> 00:09:29.993
Andy Whiteside: You know, I was listening to Elon Musk book over the weekend while I was well, I was on vacation, and they talked about as people transition from being you know, gearheads on cars to being you know, when everything became kind of sealed and you couldn't get to stuff the way used to could, where you couldn't just bolt on new stuff

83
00:09:30.465 --> 00:09:46.234
Andy Whiteside: and now we live in a world, and that's when. And the same thing happened in the PC world. And that's when we became more software centric where we got these things like these systems and these system boards. And these, you know, these chip sets. We can't really monkey with those, but we sure can write software and things to make them more efficient and and

84
00:09:46.762 --> 00:09:53.597
Andy Whiteside: more higher performing Lev levers. What's inside that box? And that's where you know Citrix. And until you when you talk, when you listen to the

85
00:09:53.960 --> 00:10:09.909
Andy Whiteside: the pundits talk about it from a financial perspective. Intel, yeah, they talk about hardware stuff and talk about it, talk about talk about it. And then they talk about the other guys trying to copy that which they are. But it's the software is where Intel has this massive head start and then take that and apply it to something like Citrix. Now you've got a the 2 working together.

86
00:10:10.040 --> 00:10:10.740
Andy Whiteside: Yeah.

87
00:10:11.856 --> 00:10:38.180
Andy Whiteside: Alright. So this first section talks about Intel Arc and Citrix remote PC access. I'll just real quick cap. Recap remote PC. For our listeners. Remote PC. Is when you use Citrix's delivery model and management model to connect to a physical PC which is gonna have a graphics card in it. They always do. But in this case a very specific graphics card that understands the what it's doing from a software and user perspective. Better, Todd, you wanna jump onto this.

88
00:10:38.970 --> 00:10:51.119
Todd Smith: Yeah, so remote. PC, was. You know that this was one of those innovations that the citrix had years ago. I mean, I I think it was probably 15 plus years ago that we had this. We introduced this.

89
00:10:51.290 --> 00:10:53.440
Todd Smith: and it was really around

90
00:10:53.720 --> 00:10:56.949
Todd Smith: customers that had invested in high-end

91
00:10:57.080 --> 00:11:03.529
Todd Smith: Pcs that were sitting in the offices. And how do you access those remote right? So it was a snow day scenario.

92
00:11:04.092 --> 00:11:11.169
Todd Smith: Or it was I need to put the physical device in a secure area and connect to it remotely.

93
00:11:11.310 --> 00:11:14.629
Todd Smith: to be able to to access it, to, to verify.

94
00:11:14.700 --> 00:11:21.662
Todd Smith: You know what are the results. And what are some of the what are some of the the images and things like that that I want to be able to to view?

95
00:11:21.960 --> 00:11:31.349
Todd Smith: So the remote PC was a really great solution for that. Marry that up with the desktop virtualization, the Vdi's type of solution. Say, well, if that machine's turned off.

96
00:11:31.420 --> 00:11:36.390
Todd Smith: We can wake it up. We can connect to it. We can have an alternative to connect to it.

97
00:11:36.420 --> 00:11:43.970
Todd Smith: But then we also can fail over back to a Vdi session or a virtual machine based solution. So it really kind of

98
00:11:44.060 --> 00:11:47.629
Todd Smith: filled out that entire delivery stack that we had

99
00:11:47.923 --> 00:11:55.289
Todd Smith: and it was all managed through the same type of policies and things like that. So there was a there was a ton of benefits towards remote PC,

100
00:11:55.668 --> 00:12:04.319
Todd Smith: when we first rolled it out. Still, to this day, we've still got a lot of good use cases being to being utilized by customers on a daily basis

101
00:12:04.590 --> 00:12:08.100
Todd Smith: leveraging this technology. But once again, it kind of

102
00:12:08.390 --> 00:12:11.890
Todd Smith: these were specific built Pcs that

103
00:12:12.090 --> 00:12:15.800
Todd Smith: you didn't want us go and have to virtualize those immediately.

104
00:12:15.930 --> 00:12:18.250
Todd Smith: Yeah. So there was. There was a lot of benefits there.

105
00:12:18.250 --> 00:12:31.450
Andy Whiteside: And that might have been because they were specialized, or it might have been. That's what you had, and you weren't ready to replace them with something else that you know cost you money again. And you know, Bill, you probably would agree that during the pandemic this this particular concept took off a lot because necessity.

106
00:12:31.960 --> 00:13:00.508
Bill Sutton: Absolutely. Yeah, this was this was definitely something that was deployed widely during the pandemic. And and actually, even before I I can remember a client in my market. What? Years ago that they gave everybody laptops cause they wanted to be able to have a have the availability to work when they were home or during a downtime event. This is before the pandemic and Cfo or the CIO walked around one afternoon, noticed how many desks still had the laptops sitting on them, so they decided to migrate to a remote PC.

107
00:13:00.800 --> 00:13:08.739
Bill Sutton: approach. So you know, another another good example of being able to leverage the technology that's sitting on the desk for access. Remotely.

108
00:13:08.740 --> 00:13:24.279
Andy Whiteside: I'll never forget the the guy in charge of disaster planning for Y. 2 K. And right there after his plan was to get rid by a laptop went forward and a VPN. Connection, and then I walked to his desk. Every day on my way out. Every day his laptop was sitting in the cradle. It did.

109
00:13:24.280 --> 00:13:26.390
Bill Sutton: Right, exactly so.

110
00:13:26.390 --> 00:13:27.030
Andy Whiteside: I need

111
00:13:27.290 --> 00:13:33.760
Andy Whiteside: so Todd anything specific around the intel Chipset here, as it relates to a remote PC, you want to call out.

112
00:13:34.030 --> 00:13:39.839
Todd Smith: Yeah, so so a lot of it is supporting. You know, everything from H. Dot 265 and AV one

113
00:13:39.960 --> 00:13:43.490
Todd Smith: support in in. It's really

114
00:13:43.680 --> 00:13:49.020
Todd Smith: the reason why that had to be done is to support a lot of the applications that are leveraging

115
00:13:49.150 --> 00:13:53.680
Todd Smith: those protocols, those those Gpu specific

116
00:13:53.690 --> 00:14:04.479
Todd Smith: initiatives, right? So everything, from being able to resolve codec issues to doing the doing the actual translations between.

117
00:14:04.510 --> 00:14:06.400
Todd Smith: you know, there's various graphic

118
00:14:06.700 --> 00:14:26.159
Todd Smith: entities that are out there and you know, a lot of it is based on some of the most popular graphics editing systems that are out there, including adobe premiere pro and you know, variety of different visual studio type of type of solutions that are out there, and

119
00:14:26.390 --> 00:14:37.400
Todd Smith: the the software that's being used, or the applications that are being used are no longer being sold to very specific users. They're they're kind of generally available.

120
00:14:37.580 --> 00:14:46.349
Todd Smith: I mean, think about it. Almost every single content Creator on Youtube uses some form of a virtual graphic studio

121
00:14:46.440 --> 00:14:48.270
Todd Smith: solution that

122
00:14:48.490 --> 00:14:57.109
Todd Smith: they're learning as they go. They're not having to go through specialized training. They're not having to buy specialized hardware to be able to do this. Yeah, they're basically going to

123
00:14:57.120 --> 00:14:59.769
Todd Smith: best buy and buying a good graphics laptop

124
00:15:01.820 --> 00:15:04.920
Todd Smith: transformed that into the business world. And now, all of a sudden.

125
00:15:05.000 --> 00:15:07.300
Todd Smith: you know, you've got the same needs.

126
00:15:07.500 --> 00:15:11.219
Todd Smith: but it has to be done more on a at scale right.

127
00:15:11.220 --> 00:15:20.288
Andy Whiteside: And that scale might be all the time, or it might be just periods of time. You don't want to over provision just off, you know, 30 min a month. Somebody's gonna need this thing.

128
00:15:21.240 --> 00:15:25.709
Bill Sutton: But you gotta. I think one of the key things here about leveraging remote PC is

129
00:15:25.780 --> 00:15:53.589
Bill Sutton: is the fact that and and this kind of goes back to the original. One of the original promises of Vdi is is keeping all of the the IP or the the really important data in the data center. So we we have a lot of customers that leverage this type of technology remote PC in particular, where they're really trying to. Maybe they have overseas or offshore folks accessing the data. And they wanna really strictly control and keep that in the data center. And by, you know, putting this on a PC. That only has access

130
00:15:53.590 --> 00:16:08.339
Bill Sutton: to you know, to what's in the data center and not allow any communication externally kind of goes back to the promise of Vdi security, centralization, etc. So you know, this certainly benefits it. And then, being able to do that for things like

131
00:16:08.676 --> 00:16:17.100
Bill Sutton: these, these high end graphics applications really makes it more effective when you start talking about things like designers or or content creators, etc.

132
00:16:17.100 --> 00:16:34.309
Andy Whiteside: Yeah, Bill, that's the number one use case I've seen over and over again. They have High End graphics needs which we're talking about here with the remote PC physical PC concept, and they had the need for centralized access to data, not only from historically the ability to to get work done, but in this day and age the ability to S. Secure the work.

133
00:16:34.310 --> 00:16:59.720
Bill Sutton: Yeah. And you know, some folks would say, Well, there's other ways you could do this without, you know, out having to bring the the Comp. Semi complex architecture to it. But you really can't, because one of the advantages here of doing this via Citrix is the Hdx protocol, hdx slash Ica protocol and the ability for that to be able to con, to understand what's going on on the endpoint and and optimize the delivery of the

134
00:16:59.720 --> 00:17:23.509
Bill Sutton: the the traffic such that the endpoint can render it in the best possible way and as quickly as possible. So the the particularly when you're talking about the arc pro here. This is a workstation a workstation gpu. So Hdx knows how to work with it and handle like Todd said. The offload of Avi and H. 2, 65, which makes it far more effective than other other methods of connectivity.

135
00:17:23.510 --> 00:17:24.109
Andy Whiteside: Right?

136
00:17:24.670 --> 00:17:48.250
Andy Whiteside: So guys, we're gonna transition into talking about virtual machines. And Vdi here, I'm gonna read from the blog this particular, this particular paragraph, because it's important. If if you're listening and you're one of those guys that when someone likes me walks in the room and touch starts talking about Gpus, and you start talking about autocad and 3D rendering. And that stuff being where this applies. Then then I'm I'm talking to you because you're wrong, cause. It applies in more places. And just that.

137
00:17:48.636 --> 00:17:57.129
Andy Whiteside: Here you go. So the benefits of Vdi with Intel Flex Data center, Gpus, traditionally, Gpus were associated with graphics, intensive workloads.

138
00:17:57.130 --> 00:18:05.029
Andy Whiteside: Only I I put the word only in there because a lot of people still think of that only. But the Vdi landscape has evolved in recent years.

139
00:18:05.050 --> 00:18:32.479
Andy Whiteside: My my point would be the last 10 years for sure. Intel Flex data center, Gpus redefine the role of Gpus in virtual desktop environments, Vdi offering enhanced performance and scalability across a variety of applications. These gpus support hardware assisted Gpu virtualization using single route I/O virtualization. Sr IV iov eliminating additional virtualization, licensing cost and complexity. Mic drop. If I would drop my mic. I'm not going to

140
00:18:33.150 --> 00:18:37.899
Andy Whiteside: that's important. And everybody needs to consider that for almost every workload.

141
00:18:37.910 --> 00:18:40.985
Andy Whiteside: anything above anything above.

142
00:18:42.153 --> 00:18:49.479
Andy Whiteside: I don't know. Task worker, and maybe even task worker needs a Gpu sliver. At least, Todd, your thoughts on that.

143
00:18:50.060 --> 00:19:01.280
Todd Smith: Yeah. So so it's an interesting. It's an interesting change in in in development here. Right? So instead of having to go and say, I need to license a very single Gpu card that I'm gonna be using here.

144
00:19:01.612 --> 00:19:16.860
Todd Smith: Whether I'm using it or not. I still need to have a license fee associated with that. That's a significant change. But I think the other piece of it is, you know, going back to the comment, you made a couple of minutes ago of of if I'm only gonna need it for 30 min

145
00:19:16.890 --> 00:19:35.679
Todd Smith: to be able to do a certain task or certain set of calculations, and then I need to put it back into the library. This is really kind of solving that solving this problem right? Eliminating the the requirement to overspend and over. Architect the solution. Here, and obviously being able to.

146
00:19:35.850 --> 00:19:44.790
Todd Smith: you know, better, deliver and reduce that total cost is huge nowadays. I think every single customer is looking for what they're spending.

147
00:19:44.930 --> 00:19:46.290
Todd Smith: what they're spending

148
00:19:46.580 --> 00:19:49.429
Todd Smith: on on for resources that they're using.

149
00:19:49.560 --> 00:19:54.229
Todd Smith: But, more importantly, what are they spending money on it for resources that they never really use.

150
00:19:54.780 --> 00:20:00.539
Todd Smith: So it's kind of a there's a there's more interest in the balance there than in the past.

151
00:20:02.130 --> 00:20:03.130
Andy Whiteside: Bill comments.

152
00:20:03.460 --> 00:20:17.266
Bill Sutton: Yeah, I would agree completely. And and it's interesting that this is license free Gpu, when it comes to, you know, carving them up into the virtual gpus. This this flex data center gpu product from Intel

153
00:20:17.899 --> 00:20:32.199
Bill Sutton: let you carve them up just like you could do with other vendors into virtual gpus and assign them to specific desktops. But to Todd's Point you had to assign a license to that, whether it was being used or not, and here it looks like it. You can just float them

154
00:20:32.200 --> 00:20:50.489
Bill Sutton: without having to worry about having the license tied up. Of course it it also, they also support the the High End Codex, just like we were talking about a few minutes ago. But here you're talking about running these in in data center server type equipment where they could be carved up and delivered very similar to the way we've done it with other products in the past.

155
00:20:50.490 --> 00:20:57.219
Andy Whiteside: Yeah, Bill's gonna ask you about that licensing concept. So explain? That's more specifically what what this changes or.

156
00:20:57.220 --> 00:21:21.889
Bill Sutton: Well in in the middle of the article based on what it says. It says Intel's license free Gpu solution for Vdi can reduce the total cost of ownership. Todd may understand it better than I do, but but at least from what I know essentially, you're not having to act. You're not having to create a license server, as it were, and have that license server say, hand out a license to each Vdi instance. This is just almost like you. You can. You can

157
00:21:21.890 --> 00:21:39.209
Bill Sutton: run as many Vgpus as the Gpu will support and you don't have to worry about the licensing so you can carve it up a number of different ways. You're not having to pay extra for the licensing worry about having a license server track it and those sorts of things, so it simplifies it, and presumably would also. Reduce the cost.

158
00:21:40.170 --> 00:21:43.630
Bill Sutton: Todd, did I catch that right? Or that's your understanding.

159
00:21:43.630 --> 00:21:51.969
Todd Smith: Yeah, yeah, you were spot on with that. It it really. It really creates a pool of resources that can be utilized as needed as opposed to

160
00:21:52.460 --> 00:21:54.290
Todd Smith: having to dedicate those right so.

161
00:21:54.290 --> 00:22:17.119
Bill Sutton: Yeah, I think, Andy, in the old, in the, in the other, a alternative ways of doing this. You actually have to assign a license to a a device, and then if and then, if you've got, or maybe a user, I I don't remember to be honest with you. And but then, if you want to add a couple of new devices in order, you had to buy more licenses, or you can just assign, move those licenses from one device for one vm to another.

162
00:22:17.421 --> 00:22:25.860
Bill Sutton: But that that's a that's a manual process that involve, you know, is prone to error. So not having that complexity in there, like I said, simplifies it.

163
00:22:25.860 --> 00:22:26.460
Andy Whiteside: Yeah.

164
00:22:26.580 --> 00:22:34.569
Andy Whiteside: yeah, going back to the original offering in the Gpu world, you basically bought a Gpu and carved it up for however many. And and it was licensed.

165
00:22:34.760 --> 00:22:38.740
Andy Whiteside: Oh, I think you actually bought an expensive Gp. There were no licenses. You just bought an expensive.

166
00:22:38.740 --> 00:22:39.930
Bill Sutton: Weren't initially. No.

167
00:22:39.930 --> 00:22:40.380
Andy Whiteside: The.

168
00:22:40.380 --> 00:23:02.700
Bill Sutton: Early days, cause I had one in a in a demo server that we had, and you just bought the gpu and you covered up. However, you wanted as as well within the bounds of the its capacity, of course, but you didn't have to worry about licensing. And then licensing was added on later in the the subsequent generation of Gpus. And and of course, here it looks like Intel's going back to that original model.

169
00:23:03.220 --> 00:23:11.889
Todd Smith: But but if you remember, when you first started doing the gpus, you had to assign the Gpus to a specific pool of resources, or a

170
00:23:11.920 --> 00:23:14.749
Todd Smith: or a site, or a farm

171
00:23:15.191 --> 00:23:26.139
Todd Smith: and they were Con. They were forever assigned to there, and you would have to carve them up based on the number of machines in your catalog that were part of that farm.

172
00:23:26.140 --> 00:23:32.719
Bill Sutton: It's much more fixed than dynamic and looks like we're we're moving over towards dynamic, which is a good thing.

173
00:23:33.100 --> 00:23:33.800
Bill Sutton: No.

174
00:23:34.130 --> 00:23:40.439
Andy Whiteside: So, guys, the the blog kind of ends with benefits to the customers. Todd, I'll let you jump in and cover this first.

175
00:23:41.190 --> 00:23:48.028
Todd Smith: Yeah. So I think the biggest benefit for the customer is, it's satisfying the need and reducing reducing those

176
00:23:49.490 --> 00:24:00.499
Todd Smith: those hurdles for adoption, or those obstacles that go in front of adoption, which is, you know, the the everyone wants the High End graphics use case. Everyone wants the the highest performing

177
00:24:00.530 --> 00:24:03.919
Todd Smith: performing device, and you know they they don't.

178
00:24:04.270 --> 00:24:11.959
Todd Smith: You know the the users will come sometimes, come up with all kinds of reasons why they want to maintain, keep their physical desktop, their high end

179
00:24:12.050 --> 00:24:14.426
Todd Smith: desktop or laptop device.

180
00:24:15.350 --> 00:24:21.747
Todd Smith: but from an from an organizational perspective it helps reduce that overall cost

181
00:24:22.260 --> 00:24:24.290
Todd Smith: whether it be from the licensing

182
00:24:24.300 --> 00:24:34.160
Todd Smith: savings from the hardware resource savings that could be out there? And and the big thing is, is, you know, the the benefits can be

183
00:24:34.420 --> 00:24:57.879
Todd Smith: can be proven right. So you leverage tools like, you know, Login, Vsi and some of the other solutions that out there that they can help provide those case studies provide the that empirical information. It's really gonna show you that. Yes, you can save more by doing it this way than if you have to go out by a fleet of High End graphics workstations as an example.

184
00:24:57.880 --> 00:25:01.699
Andy Whiteside: And then those get broken. Those get stolen. There's security concerns.

185
00:25:02.210 --> 00:25:04.739
Andy Whiteside: There's data. Synchronization concerns

186
00:25:05.010 --> 00:25:05.530
Andy Whiteside: the.

187
00:25:05.530 --> 00:25:18.020
Todd Smith: Well and and quite frankly, you know, one of the other challenges with using a a laptop with a high end. Graphics card is, if you've got a sitting on your lap for a while your your legs start cooking.

188
00:25:18.530 --> 00:25:19.090
Bill Sutton: Yeah.

189
00:25:21.030 --> 00:25:28.920
Andy Whiteside: So, Bill, take this conversation. But now twist it over towards the knowledge worker and even the task worker. And how supplies in the Vdi world for benefits.

190
00:25:29.110 --> 00:25:58.710
Bill Sutton: Yeah, I think when you start looking at the knowledge worker in particular. Well, really, both of them, it it goes back to what I was saying earlier around leveraging standard applications today, and the benefits of having Gpus are gonna accelerate the performance to those applications like again, like Excel or Powerpoint in particular, because of the math and the calculations that go into rendering and calculating everything in those in those tools. So certainly. There's benefits to customers there and the but the key thing about

191
00:25:58.710 --> 00:26:18.100
Bill Sutton: these this new generation of Gpus, and it's in the blog article itself, this choice and flexibility. You know, you've got license free options. You've got all support for all the codecs. All of these things will help improve the experience for the end user even the task worker and the knowledge worker particularly when they're leveraging

192
00:26:18.100 --> 00:26:30.026
Bill Sutton: a standard office app, so that whether you know it or not. If you've if you've got a a a standard or relatively recent laptop today, you can look in task manager and see the Gpu graph and it, and it

193
00:26:30.340 --> 00:26:41.399
Bill Sutton: it actually goes up a lot when you're doing things in basic things in Powerpoint, in excel. So having that capability on the end endpoint. I'm sorry in a Vdi environment is key.

194
00:26:41.400 --> 00:26:45.659
Andy Whiteside: And the one we haven't talked about, which I constantly go look at in my world is just using a browser.

195
00:26:45.820 --> 00:26:47.447
Bill Sutton: Yeah, exactly. Exactly.

196
00:26:47.990 --> 00:26:56.651
Andy Whiteside: And that's where you're gonna see tab after tab after tab, just sucking up the Gpu resources if it has them, which in a physical world. It definitely has them.

197
00:26:56.940 --> 00:27:02.879
Bill Sutton: Let's talk about key Gpu resources. Let's talk about what we're doing right now, you know. So.

198
00:27:04.622 --> 00:27:10.167
Andy Whiteside: Okay. Well, guys, thanks for jumping on and covering this topic. This is a relevant topic, and it will continue to be

199
00:27:10.510 --> 00:27:15.799
Andy Whiteside: for as long as the end user compute world exists, I'm sure. We'll keep an eye on it for you guys.

200
00:27:15.800 --> 00:27:17.650
Bill Sutton: For sure. Thanks, Andy.

201
00:27:17.650 --> 00:27:20.000
Andy Whiteside: Alright guys talk to you next. Thank you.