[00:00:00] Speaker 01: And we'll hear the next argument, case number 22-1038, Intel Corporation against PACT-XPP Schweizer AG. [00:00:09] Speaker 01: Mr. Mammon. [00:00:12] Speaker 02: Good morning, and may it please the court. [00:00:14] Speaker 02: Nathan Mammon on behalf of Intel Corporation. [00:00:18] Speaker 03: The board's decision upholding claim four of the 631 patent should be reversed for two independent legal errors. [00:00:25] Speaker 03: First, the board adopted an illegally erroneous claim construction of the claim term plurality of bus segments that limited them to only directly connected bus segments. [00:00:36] Speaker 03: And based on that erroneous construction, concluded that the Bozinski reference did not disclose those bus segments. [00:00:43] Speaker 03: And second, the board improperly faulted Intel for not specifically explaining how Arameli's shared bus structure would be physically incorporated into King. [00:00:51] Speaker 03: Either Arameli warrants vacating or outright reversing the board's decision in this case. [00:00:56] Speaker 00: We've all got a lot of things going on here and a lot of claims. [00:01:01] Speaker 00: This is just a process question, though. [00:01:03] Speaker 00: I understood that in your briefing, what you were saying is if we reverse the board on [00:01:10] Speaker 00: combination of King and Erroll Milley that you were advocating for reversal. [00:01:15] Speaker 00: But I thought you were only advocating for a remand on the issue of the corality of buses. [00:01:21] Speaker 03: That is correct, Your Honor. [00:01:22] Speaker 03: So on the claim construction point, if the court reverses the board on the claim construction, that would at minimum require [00:01:28] Speaker 03: remand for consideration of the Budzynski and Arameli combination, which the board rejected on the clean construction. [00:01:36] Speaker 00: But you already also make an argument up here about the combination of the Budzynski, whatever reference to, right? [00:01:43] Speaker 03: There are arguments, yes, but even under the board's construction that we submit that even under the board's erroneous construction, that Budzynski does show those directly connected bus segments that comprise a plurality of flexible data paths. [00:01:58] Speaker 03: So there is an argument that we made that we think the board's construction was wrong. [00:02:03] Speaker 03: But even under that construction, Budzinski and Arameli would still show that. [00:02:07] Speaker 03: But the board rejected the argument of Budzinski on claim construction. [00:02:16] Speaker 03: So I think for that ground, the natural thing might be to remand for the board to consider the combination that it didn't actually address. [00:02:23] Speaker 03: On the King-Arameli ground, though, however, [00:02:25] Speaker 03: If the board applied an erroneous legal standard for obviousness and under the correct legal standard, which would not require bodily incorporation, that should be a reversal. [00:02:35] Speaker 03: There's really no dispute, we submit, that the board found for claims two and three that you would combine King and Air Millings. [00:02:44] Speaker 00: So let's talk about bodily incorporation. [00:02:46] Speaker 00: Because bodily incorporation is just a subset of motivation to combine, right? [00:02:51] Speaker 00: It's not, I mean, doctrinally, it's not its own sort of separate doctrine. [00:02:56] Speaker 00: It's just an aspect, a subset of motivation to combine. [00:03:01] Speaker 03: It can be, Your Honor. [00:03:02] Speaker 03: I think, though, here that the board's focus was not on whether there was motivation, whether Aera Milley would improve the King system. [00:03:13] Speaker 03: The board here focused on, could you actually combine Aera Milley into King? [00:03:18] Speaker 03: And in fact, the board pointed to the wrong part of Aera Milley. [00:03:21] Speaker 03: So at Appendix 38, the board faulted Intel for not explicitly explaining how to incorporate the shared bust as disclosed in Arameli. [00:03:29] Speaker 03: That was never our argument below. [00:03:31] Speaker 03: Our argument below was Arameli teaches tertiary cash, teaches the advantages of tertiary cash to speeding a processor. [00:03:38] Speaker 03: You would apply that teaching of tertiary cash into the King system. [00:03:42] Speaker 03: The board faulted us for saying that. [00:03:43] Speaker 00: But there are limits, as we just discussed in the earlier case. [00:03:46] Speaker 00: There is Judge Hughes' rights and our case law. [00:03:50] Speaker 00: If it's inoperable, then it feels under motivation to combine. [00:03:55] Speaker 00: Or at least we have certain cases that say that under the facts of those cases. [00:04:00] Speaker 00: So is bodily incorporation kind of the cousin of something being inoperable? [00:04:08] Speaker 03: I don't think it necessarily is. [00:04:11] Speaker 03: And in this specific case, I don't think there was a dispute that somehow this would be inoperable. [00:04:17] Speaker 03: Pax expert below, and I can point you to appendix 3514 to 3515 and other places, wasn't that the circuit wouldn't work. [00:04:26] Speaker 03: Wasn't that you couldn't combine these. [00:04:28] Speaker 03: It was that you'd have to make modifications. [00:04:31] Speaker 03: That was the phrase that was repeated. [00:04:33] Speaker 03: throughout PACS evidence was you'd have to make modifications. [00:04:36] Speaker 03: You'd have to make modifications. [00:04:37] Speaker 03: It wasn't that those modifications were beyond the skill of a person of skill in the art. [00:04:41] Speaker 00: Well, did the petition allege or did you allege that those modifications are not outside of one's skill in the art? [00:04:48] Speaker 03: Well, certainly we did. [00:04:50] Speaker 03: And I pointed to both in the petition, Appendix 200, Appendix 202, which was explaining how a person of skill would combine the L3 caches with King's bus system. [00:05:03] Speaker 03: Our expert explained in Appendix 783 that a person would know how to make this arrangement. [00:05:08] Speaker 03: And in his deposition in Appendix 3604-05, [00:05:12] Speaker 03: He provided further elaboration and under questioning from Pack about how a skilled person with skill in the art would know to make the combination of the tertiary cash as taught in Aramillie into the King system. [00:05:24] Speaker 03: And so we provided the motivation that the tertiary cash would improve the operation of King as King is looking for to improve processor speed that Aramillie's [00:05:36] Speaker 03: Tertiary cash, shared cash would also do that. [00:05:39] Speaker 03: We explained that person of skill would know how to do it. [00:05:42] Speaker 03: Pax evidence was not that this could not be done. [00:05:44] Speaker 03: That would be inoperable. [00:05:46] Speaker 03: Pax evidence was you'd have to make modifications to do that. [00:05:48] Speaker 03: And this court's precedent doesn't ally directing in other cases. [00:05:53] Speaker 03: The point is not whether there's need to be modifications. [00:05:55] Speaker 03: It's whether it would be obvious to a person of skill to do it. [00:05:58] Speaker 03: But if I may, where the board, I think, most clearly went wrong, again, at appendix 38, [00:06:03] Speaker 03: is that it was even looking at the wrong thing to change. [00:06:07] Speaker 03: And I think this is a telltale indication of where they went astray on bodily incorporation. [00:06:13] Speaker 03: But they were requiring an explicit explanation of how to incorporate the shared bus as an air mill. [00:06:18] Speaker 03: And that wasn't the argument of the combination that was being made below. [00:06:22] Speaker 03: The argument in the petition was you'd incorporate the tertiary cash, and you would use that in the King system. [00:06:30] Speaker 03: And so the board was focusing on basically setting up this false comparison of, hey, Arameli has a shared system, but King uses this direct-to-direct connection, and those are incompatible. [00:06:47] Speaker 03: But that was never the argument. [00:06:51] Speaker 03: And beyond that, the board also just fundamentally misread Aramilla, that Aramilla's teaching of how to use a tertiary cash wasn't limited. [00:06:59] Speaker 00: So your argument on motivation to combine isn't just they legally erred by requiring incorporation. [00:07:07] Speaker 00: It's they were just looking at the wrong things in the wrong way. [00:07:11] Speaker 03: I think it's both. [00:07:13] Speaker 03: I mean, they legally error by requiring incorporation of the wrong thing. [00:07:17] Speaker 03: They were focused on, and I think that's sort of indication of the board's body incorporation error. [00:07:22] Speaker 03: They were focused on taking Aramilly, its shared bus, and cash, and how do you put that into the King system, rather than focusing on the teaching of Aramilly, on the cash, shared cash teaching, and how would that be incorporated into King. [00:07:37] Speaker 03: And so I think they're certainly overlapping, slightly different in the focus on the wrong thing, on the shared bus, but also improperly requiring incorporation of that shared bus into King. [00:07:51] Speaker 00: Can you just move to that? [00:07:53] Speaker 00: Unless my colleagues have anything on this point. [00:07:55] Speaker 00: And this case also presents a direct versus indirect connection question, right? [00:08:01] Speaker 00: Yes, sure. [00:08:03] Speaker 00: And was the board clear here that it was requiring a direct connection? [00:08:08] Speaker 03: I think it's crystal clear. [00:08:09] Speaker 03: If you look at appendix 18, the board construed plurality of bus segments to mean, quote, that each bus segment is directly accessible to its corresponding processor or processors. [00:08:21] Speaker 03: And then a couple of pages before that, appendix 16, the board said a plurality of bus segments for each processor does not include bus segments that are only indirectly accessible to a processor. [00:08:32] Speaker 00: So are we supposed to view this as a claim construction question, or as a substantial evidence question, as it arises here? [00:08:39] Speaker 03: So on the Bozinski ground and on the claim construction, it is, at first point, a claim construction issue, because the board construed [00:08:48] Speaker 03: this claim term in a way that's fundamentally contrary to the claim language and the intrinsic record. [00:08:54] Speaker 00: OK, so let's say hypothetically we were to conclude that the board was wrong and that it doesn't require direct connection. [00:09:01] Speaker 00: It can be in connect. [00:09:02] Speaker 00: How far does that get us in terms of resolving this case? [00:09:05] Speaker 03: So it gets you to, we say that if the board's wrong, that it does not require a direct connection, that Bozinski shows those connections. [00:09:14] Speaker 03: PAC did not dispute that in their briefing, that Bozinski would show that. [00:09:18] Speaker 03: So the limited work for the board to do on remand would be to address, OK, Bozinski shows everything in claim one. [00:09:26] Speaker 03: Now, address the arguments that were made on motivation to combine Bozinski with Erroll Milley. [00:09:31] Speaker 03: So that would be the work for the board on remand would just be to address the motivation combined for the Budzinski-Aremilli combination, which they didn't reach because they decided this on the claim construction term. [00:09:45] Speaker 03: But again, with the King-Aremilli, that's independent of the claim construction issue. [00:09:49] Speaker 03: And the board's heir of Bottling Corporation, I think, is something that [00:09:56] Speaker 03: They didn't address in the Bozinski part, but the art is different. [00:10:01] Speaker 03: I think that it wouldn't. [00:10:05] Speaker 03: The board's Aaron King, I guess, didn't permeate to Bozinski because the board never actually addressed combining King, excuse me, combining Bozinski and Aaron Mellich. [00:10:17] Speaker 03: But if I may just ask the question that I'd go back to the claim construction point, where I think so clearly the board went wrong is that the plurality of bus segments must comprise a plurality of flexible data channels. [00:10:33] Speaker 03: And there's no dispute that the flexible data channels are created by switches. [00:10:39] Speaker 03: That's actually found, you can see it in the figure that's on page 16 of a reply brief. [00:10:45] Speaker 03: which is actually a figure that came from Pat's brief as well at page 39. [00:10:50] Speaker 03: But the figure on page 16 of our reply brief is a figure six from the patent. [00:10:56] Speaker 03: And here, these are annotations by Pat. [00:11:00] Speaker 03: They show the green boxes. [00:11:02] Speaker 03: They show those directly connected bus segments. [00:11:05] Speaker 03: And in the blue boxes, they show, well, these are the indirectly connected bus segments. [00:11:10] Speaker 03: But the only way you get the flexible data channels is with those indirect connected bus segments. [00:11:17] Speaker 03: And so that's how this [00:11:20] Speaker 03: works in context of the 6-1 pattern, but if you look at the claim language, the bus segments, plurality of bus segments, must have these flexible data channels, which ergo it must have at least these indirect connections to meet the claim. [00:11:38] Speaker 03: I think where the board went astray is treating a plurality of flexible data [00:11:42] Speaker 03: excuse me, a plurality of bus segments as a separate limitation from the plurality of flexible data channels, rather than understanding it as it properly is in the claim that the flexible bus segments must have the flexible data channels, that it's one limitation that the flexible data channels further defines what is required by those bus segments. [00:12:09] Speaker 03: And so for that reason, we think the board's construction is wrong, and that it would require a reversal of the King, excuse me, of the Dzinski and Aramilly grounds. [00:12:20] Speaker 03: But we think the court can reverse outright on King and Aramilly. [00:12:26] Speaker 03: Unless the court has further questions, I'd be happy to reserve my time for rebuttal. [00:12:30] Speaker 01: Any questions at this moment? [00:12:33] Speaker 01: Thank you. [00:12:35] Speaker 01: Mr. Weisberg. [00:12:36] Speaker 05: May it please the court, Sanford Weisberg for Pelley Pact again. [00:12:39] Speaker 05: I'd like to begin with this question of bodily incorporation and specifically Appendix 38. [00:12:45] Speaker 05: And I think really what the bodily incorporation argument is is seizing upon what we believe clearly is a typo in the board's decision at Appendix 38. [00:12:54] Speaker 05: And I'd like to explain that for a minute. [00:12:57] Speaker 05: The supposed error that the board made in misconstruing Intel's arguments was it said, well, you need to show how the shared bus from Aramilly will be used in your combination. [00:13:08] Speaker 05: And the board, indeed, that was a typo. [00:13:10] Speaker 05: The board meant to say a shared cache, not a shared bus. [00:13:15] Speaker 05: And I'd like to explain why we know that that's the case. [00:13:18] Speaker 05: If you look at the second sentence on appendix 38, the board correctly characterizes Intel's argument as the combination of King and Arameli in which a shared cache retains the 16 separate directly connected bus segments as shown in King. [00:13:34] Speaker 05: No reference to a shared bus. [00:13:36] Speaker 05: That was a correct interpretation of Intel's argument there. [00:13:40] Speaker 05: In the subsequent sentence that begins [00:13:43] Speaker 05: that arbitration was a well-known bus mechanism, the board mistakenly referred to a shared bus. [00:13:49] Speaker 05: Now, how do we know this additionally, besides the fact that the board got it right in the first sentence, is that the board cites in this second sentence the patent owner's surapply at 11 to 12, which is also in the appendix, and that is at page 576. [00:14:06] Speaker 05: At those pages of the patent owner's surpli, there's no reference to a shared bus. [00:14:10] Speaker 05: There's only a reference to the shared cash being used with King's bus segments. [00:14:15] Speaker 05: So one typo by the board does not warrant reversal or vacatur. [00:14:20] Speaker 05: This court, previously in a case involving the Trademark Board, found that there was a typo. [00:14:24] Speaker 05: That's the Omaha Stakes case, 908 F3rd at 1327, and did not find a cause for reversal on that point. [00:14:32] Speaker 05: So, coming back to this bodily incorporation argument, if one understands that this is indeed a typo, the board did understand Intel's combination. [00:14:43] Speaker 05: Again, as involving, let's take the shared cash from Aramilli, let's combine it with the 16 bus segments from King. [00:14:49] Speaker 05: That was their own combination. [00:14:51] Speaker 05: It's at Appendix 202, which is the petition. [00:14:53] Speaker 00: Can I just ask you, did you make this typo argument in your brief? [00:14:56] Speaker 05: We didn't use the word typo, but what we did say, but we did, what we did argue is that the board understood the Intel's argument involved using the shared cash from Aramilly. [00:15:08] Speaker 01: Yes, that makes sense. [00:15:10] Speaker 01: But we still need to understand [00:15:13] Speaker 01: why that would change the result. [00:15:15] Speaker 05: Yes, absolutely. [00:15:16] Speaker 05: So if I could turn to the question. [00:15:19] Speaker 00: But I'm sorry. [00:15:20] Speaker 00: I'm just a little, you make an argument about a typo, which sounds like a compelling argument that if you just misunderstand it. [00:15:29] Speaker 00: And it's the first time I'm hearing about a typo here. [00:15:33] Speaker 00: So show me in your group where you covered this issue, even if you didn't use the word typo, make me feel better. [00:15:42] Speaker 05: First point I would make is one of the main focuses on our brief was a different combination that was proposed in the reply involving a crossbar. [00:15:48] Speaker 05: So a lot of discussion on that. [00:15:51] Speaker 05: Then after that, we turned to the question of the original petitions proposed combination, which was the 16 bus segments from King combined with the shared cash, not the shared bus of Aramilli. [00:16:06] Speaker 05: And if you just bear with me for one moment, [00:16:14] Speaker 05: OK, so page 36 of our brief, the red brief, before quoting the same paragraph on page 38 of the appendix that I was just talking about, we say, the board also addressed Intel's original proposed combination, i.e. [00:16:30] Speaker 05: that King's 16 separate bus segments could be modified to include Aramilly's L3 caches. [00:16:36] Speaker 05: That is how we interpreted the board. [00:16:38] Speaker 05: Now, I will admit, we did not use the word typo. [00:16:42] Speaker 05: We did not specifically dissect that last sentence of the board's paragraph at appendix 38. [00:16:48] Speaker 05: But our brief did argue that the board addressed. [00:16:50] Speaker 00: It's a somewhat different argument. [00:16:52] Speaker 00: And it's on a key point here. [00:16:54] Speaker 00: I mean, you seem to be saying, [00:16:55] Speaker 00: You started off, if I'm recalling what happened five minutes ago, that you said, this is not a bodily incorporation case, because all it's built on is a sentence which is a typo in the board's thing. [00:17:10] Speaker 05: So if I could move on to the question of Intel's proposed combination in the petition about the L3 cash being combined with King's 16 separate bus segments, the board did address that in that paragraph and I'd like to explain why the board found that there was not going to be a motivation to combine. [00:17:27] Speaker 05: And I think it's helpful here to look at a picture, because it's a lot of words. [00:17:32] Speaker 05: But if we look at a picture of appendix 26, this is the board's decision again, you see King. [00:17:38] Speaker 05: And you see the 16 bus segments in King. [00:17:43] Speaker 05: What Intel was proposing to do was to preserve that bus system, but to attach these caches to each of those bus segments. [00:17:54] Speaker 05: That's the proposal, it's not my version, it's not the board's version, it's Intel's own proposed combination in Appendix 202, and I'll read from it again, is that they're taking the L3 caches taught by Arameli, and we're going to directly attach them to King's processors P0 through P3, and connect them over King's processor buses. [00:18:14] Speaker 05: Now, here's what the problem with that. [00:18:17] Speaker 05: Now, is it physical inoperability? [00:18:19] Speaker 05: Perhaps not, but it requires substantial modification, because what you need, if you're going to accomplish that combination, you need to have communication between all of these 16 processor buses. [00:18:30] Speaker 05: in a way that King does not require. [00:18:32] Speaker 05: King doesn't require communication, because in King you've got a direct line from each processor to each memory. [00:18:37] Speaker 05: Now that we're sharing a cache memory among all these P0 through P3 processors, we need some way for cache coherence, as well as for other purposes, of communicating between these 16 processor buses. [00:18:49] Speaker 05: And Intel says, well, arbitration. [00:18:51] Speaker 05: That's how you do it. [00:18:52] Speaker 05: But the board had substantial evidence to say, this is not just throwing up the word arbitration. [00:18:59] Speaker 05: It's a complicated mechanism to enable that communication. [00:19:02] Speaker 05: Not only that, but the board went a step further at Appendix 38. [00:19:05] Speaker 05: And it said, not only is it complicated, but the whole point of King is to avoid competition and latency in terms of a processor accessing a memory. [00:19:15] Speaker 05: Once we introduce this idea of communication, we're going to have a lot of waiting around, a lot of competition for who's using which bus when. [00:19:21] Speaker 05: And the board had substantial evidence to find that the skilled artisan would not do it. [00:19:25] Speaker 05: I'd like to return to this physical inoperability point. [00:19:28] Speaker 05: It came up in the prior case, but it's also clearly important. [00:19:32] Speaker 05: Again, our position, respectfully, is that the patent owner's burden is not to show that the combination fails because it's physically inoperable. [00:19:40] Speaker 05: Could that be a way? [00:19:41] Speaker 05: Yes. [00:19:42] Speaker 05: But what the patent owner is allowed to show is that there are substantial obstacles and costs to doing it. [00:19:49] Speaker 05: Don't take my word from it for it. [00:19:52] Speaker 05: Please take a look at your honor's discretion at the Henny Penny case. [00:19:55] Speaker 05: And I'd like to just discuss the facts very briefly. [00:19:58] Speaker 05: Henny Penny was a patent on a [00:20:01] Speaker 05: detecting whether there were impurities in frying oil. [00:20:06] Speaker 05: And they were trying to combine two patents there. [00:20:09] Speaker 05: And the combination, one of the patents, the one that actually existed for impurities infused in crude oil, cooled the oil down before it tested for the impurities. [00:20:18] Speaker 05: And what the combination wanted to do was, let's not cool the oil down. [00:20:21] Speaker 05: Let's test it when it's really hot. [00:20:23] Speaker 05: And they said, OK, well, that might mean that our sensor doesn't last as long. [00:20:28] Speaker 05: And that was the cost of the combination. [00:20:31] Speaker 05: Our sensor is not going to last as long. [00:20:33] Speaker 05: And this court said, OK, there's a benefit, which is you don't have to cool the oil down. [00:20:39] Speaker 05: But there's a cost, which is your sensor might not last very long. [00:20:42] Speaker 05: And there is substantial evidence, this court found, for the board to say, in light of those costs and benefits, that a skilled artisan would not have made that combination. [00:20:50] Speaker 05: It was not a physical and operability question. [00:20:52] Speaker 05: It was at cost and benefit. [00:20:54] Speaker 05: And just to refer briefly to the exact page, well, it's in our brief. [00:20:59] Speaker 05: But it's basically saying you can't just say arbitration. [00:21:03] Speaker 05: You can't just say that you're going to incur a cost, and that's OK. [00:21:09] Speaker 05: You have to weigh the cost and the benefit. [00:21:10] Speaker 05: The skilled artisan has to do that. [00:21:12] Speaker 05: The board has substantial evidence to look at what a skilled artisan would do. [00:21:15] Speaker 05: And so here, coming back to my picture at appendix 26, you've got all these 16 bus segments. [00:21:21] Speaker 05: You're going to need communication between them. [00:21:23] Speaker 05: The whole point of King is not to have contention and competition between the processor getting to its memory. [00:21:29] Speaker 05: And the board had substantial evidence to say, well, that combination that Intel is proposing is going to be too problematic. [00:21:36] Speaker 05: And a skilled artisan would not be motivated to make it. [00:21:39] Speaker 05: If I could turn to the claim construction plan. [00:21:42] Speaker 05: A couple of points. [00:21:43] Speaker 05: One is that the claim language is a plurality of bus segments for each processor. [00:21:49] Speaker 05: We then get to this flexible data channels point, but that's after. [00:21:52] Speaker 05: For each processor, the meaning of four, dictionary definition, is belonging to. [00:21:58] Speaker 05: If you reach out and you have all these indirectly accessible segments elsewhere in the diagram in figure six, and if those count as four, then it deprives four of any meaning. [00:22:08] Speaker 05: Even better, though, than the dictionary definition is that Appendix 1798 from the file history, the very claim language that we're talking about was amended. [00:22:17] Speaker 05: It went from plurality of segments to plurality of bus segments for each processor. [00:22:22] Speaker 05: This was an amendment that PAC had to make in order to get its patent approved. [00:22:27] Speaker 05: And what it did was it pointed at this as at Appendix 1798, [00:22:31] Speaker 05: It pointed to the figure six from its patent, and it pointed specifically to which bus segments it was talking about when it was adding these words for each processor. [00:22:42] Speaker 05: And the ones it pointed to was the four input direct segments going into a processor and the four going out of the processor, eight in total. [00:22:49] Speaker 05: If it had meant to incorporate into this other indirectly accessible ones, it would have pointed to those, but it did not. [00:22:56] Speaker 05: And that, we believe, is the most telling explanation. [00:22:59] Speaker 05: Now, Intel says, well, they use the word example, but that's respectfully a red herring because [00:23:05] Speaker 05: It was an example that in that it only involved four in that case But what it was interpreting in that diagram was which ones count Okay, if it had chosen a different diagram might have been a different number But in this diagram that it chose an example if it had meant to include the indirect ones It would have pointed to those and what Intel is basically doing is it's saying interpret the claim language before it was amended [00:23:28] Speaker 05: When it just said plurality of segments full stop, they're ignoring and not giving any meaning to the words plurality of bus segments for each processor. [00:23:37] Speaker 00: But you're not discussing the additional words in the claim, which are flexible data channels. [00:23:42] Speaker 00: I'd like to do that. [00:23:44] Speaker 00: Thank you. [00:23:44] Speaker 00: Thank you, Judge Frode. [00:23:45] Speaker 00: Because I think that's the argument. [00:23:47] Speaker 00: Yes. [00:23:47] Speaker 05: So the claim goes on to say comprising flexible data channels to each processor. [00:23:52] Speaker 05: And they argue two is broader than four, et cetera. [00:23:56] Speaker 05: Here's the problem with that. [00:23:58] Speaker 05: As I heard their argument, and as I also understood it from their brief, what they're saying is pact has to show that there would be an element to element path in order for it to be a flexible path. [00:24:08] Speaker 05: And you can't do that without using the indirect ones. [00:24:12] Speaker 05: But the patent doesn't say it has to be element to element. [00:24:16] Speaker 05: And the fact that this language in isolation about flexible data channels to each processor is broader in isolation from the plurality of bus segments for each processor, the way to reconcile that, we submit, is the fact that the limiting factor is what comes first. [00:24:33] Speaker 05: Plurality of bus segments for [00:24:36] Speaker 05: comes before comprising flexible data channels, too. [00:24:39] Speaker 05: And therefore, if there's a limitation that narrows it in the first part, which we think there is because of the word for and because of the file history that I referenced. [00:24:46] Speaker 00: So are you saying these words don't mean anything? [00:24:49] Speaker 00: No, no. [00:24:49] Speaker 05: They do mean something. [00:24:51] Speaker 05: And in fact, it comes up later in this Ludzinski combination. [00:24:54] Speaker 05: What it means is that you need more than one path to send data at the same time. [00:24:59] Speaker 05: Okay, so you don't need to send it all the way to the next element, but you need some way, if you're talking about a processor, you need at least two ways of getting data into and or out of that processor at the same time. [00:25:11] Speaker 05: That's what flexible data channels requires. [00:25:13] Speaker 05: And the reason why this Budzinski-Eramilli combination goes wrong, one of their arguments, is that [00:25:20] Speaker 05: They try to use a status bus and a data bus to be the two segments. [00:25:25] Speaker 05: And the status bus, even if it constitutes data, which the board found it didn't, is using an entirely different sort of signal than what's going through the data bus. [00:25:34] Speaker 05: And so it's not flexible, because you don't have the option of using both at the same time. [00:25:39] Speaker 05: They're for different purposes. [00:25:41] Speaker 05: And so that's what flexible data channels means. [00:25:43] Speaker 05: It requires you to have more than one option. [00:25:46] Speaker 05: But it doesn't require you to go from an element all the way to another element over a single bus segment. [00:25:52] Speaker 05: And again, plurality of bus segments for each processor is the key language that comes first and imposes this limiting restriction. [00:25:59] Speaker 05: I addressed status bus briefly. [00:26:03] Speaker 05: The other argument that Intel makes has to do with one segment in Budzinski, which it says, it said in its petition at Appendix 209 that that was one segment. [00:26:13] Speaker 05: Now it's saying, well, no. [00:26:15] Speaker 05: And in its reply at the board, it said for the first time, these are two segments, two sub-segments, if you will. [00:26:20] Speaker 05: The board found that this was a new argument on reply. [00:26:23] Speaker 05: That was within its discretion not to address it. [00:26:24] Speaker 05: But in any event, the board had substantial evidence to reject it on the merits. [00:26:28] Speaker 05: And the reason is because that argument is entirely predicated on one sentence from Butzinski, which is an extremely long patent on specification. [00:26:35] Speaker 05: And what that sentence says is the word, it uses the word contemporaneous data transmission. [00:26:41] Speaker 05: That's Appendix 2129. [00:26:43] Speaker 05: The board had substantial evidence to find that contemporaneous does not necessarily mean simultaneous. [00:26:49] Speaker 05: Of course, in the computing world, things are happening very quickly, one thing on top of the other. [00:26:54] Speaker 05: But it would have to be simultaneous to be able to prove that this were two flexible bus segments. [00:27:00] Speaker 05: And Buzinski does not carry that burden for them. [00:27:03] Speaker 05: And therefore, the board had substantial evidence to reject that combination as well. [00:27:08] Speaker 05: If the court has no further questions, I'll rest on my brief and ask that the board be affirmed in this case. [00:27:15] Speaker 01: Thank you. [00:27:15] Speaker 01: Thank you. [00:27:20] Speaker 03: Thank you, Your Honor. [00:27:21] Speaker 03: So if I may start with the argument that we heard, which was the first time that this was regarding to the King and Aramello accommodation, that this was a typo. [00:27:29] Speaker 03: It wasn't an argument that had been made before in the briefs. [00:27:34] Speaker 03: But it also, I don't think, is a fair reading of the board's decision or the record. [00:27:39] Speaker 03: And you can look at Appendix 35 of the board's decision. [00:27:42] Speaker 03: where the board, in its process of describing what each side is arguing, they note that Intel is saying that PAC is requiring a combination to bodily incorporate Paramillily's bus into King's system, which Intel denies is necessary. [00:27:55] Speaker 03: So the board recognized that that was what Intel was saying PAC was requiring. [00:28:00] Speaker 03: And it wasn't just an Intel argument. [00:28:02] Speaker 03: You go look at PAC's own submissions to the board at appendix 461, which was PAC's response. [00:28:09] Speaker 03: To our petition they say that it wasn't possible to It's pointed to there are merely l3 caches relying on the shared bus system and [00:28:23] Speaker 03: of Aramilli. [00:28:25] Speaker 03: Their expert also made this point at appendix 3513 that, again, Aramilli had the shared box. [00:28:31] Speaker 03: So that was the argument that PAC had made to the board. [00:28:34] Speaker 03: When the board addresses that argument on page 38 of its decision, it was addressing the argument that PAC had actually made. [00:28:42] Speaker 03: But even if you think that this was a typo, I would have two responses on that. [00:28:46] Speaker 03: That one, it's the key decision or key language in the board's decision rejecting Intel's arguments. [00:28:52] Speaker 03: At the very least, on this key language where the court requires reasons, decision-making from the board, the board should be required to clarify that, that that actually was a typo, because we don't think it actually was. [00:29:05] Speaker 03: But even sending that report, whether there's a remand required just for the board to clarify its decision, at the very least, this is still a requirement of bodily incorporation, even if you don't think that shared busts [00:29:16] Speaker 00: What about the Henny Penny case, which they would certainly cover it in their brief. [00:29:23] Speaker 00: And they cited as saying that if the board is entitled to rely on expert testimony, that the combination would add additional complexity and lead to decreased efficiency. [00:29:36] Speaker 00: So is that something that was shown here by substantial evidence, which is a problem for you? [00:29:43] Speaker 03: So I don't think it's a problem for us, Your Honor, that the Hennepinney case was a decide on the motivation to combine. [00:29:49] Speaker 03: And the board, in that case, in the court, affirmed the board's decision that, look, because of these differences in the art, there wouldn't be the motivation to combine. [00:29:57] Speaker 03: I know we had the earlier exchange of sort of the relationship of motivation to combine with bothering corporation. [00:30:02] Speaker 03: But here, I think the problem is the board was not focused on [00:30:06] Speaker 03: what the teaching was that was being combined. [00:30:09] Speaker 03: So they never actually got to the point that the board did in Henny Penny of saying that, hey, you couldn't combine the Arameli cash with Keen. [00:30:18] Speaker 03: And so we think that Henny Penny just was dealing with a separate issue. [00:30:23] Speaker 04: I'm a little confused, but let me see if this is. [00:30:26] Speaker 04: Your view is that the board's error is that it actually did require [00:30:32] Speaker 04: these two references to be physically combined, and that's legal error. [00:30:37] Speaker 04: And that's why it rejected a motivation to combine. [00:30:40] Speaker 04: I think your friend's arguing a more subtle way of saying, well, it's not that they can't physically be combined, but when you look at these two references to achieve the goals of the patent, a skilled artisan would not know how to combine them. [00:30:55] Speaker 04: And because there's a lot of complicated factors that would have to be in. [00:31:01] Speaker 04: And so they wouldn't look to these two references. [00:31:04] Speaker 04: I don't think I've articulated it very well, but it seems those are two different arguments. [00:31:11] Speaker 04: And you want me to read the board's decision as the bodily incorporation, and he wants me to read it as that other kind of, well, a skilled artisan wouldn't do this, not because of bodily incorporation, but because [00:31:23] Speaker 04: a skilled artisan would not understand that these could be combined, because there's so many other things you would have to do. [00:31:31] Speaker 04: Is the second way a fair reading of the board's decision in your view? [00:31:35] Speaker 03: I don't believe it is, because all we have to go on is what's on Appendix 38, and what the board's rationale was, and what the board was saying that hadn't been shown was, how do you incorporate the Aramilli shared bus? [00:31:46] Speaker 03: into King's system. [00:31:47] Speaker 03: So the starting point wasn't even how, if anything, if there's an argument that, well, the person wouldn't physically combine, it was because, well, the board was looking at the wrong thing to combine. [00:31:58] Speaker 00: So if we sent it back and the board clarifies and they say, on this record, you know the record better than I, [00:32:05] Speaker 00: that we didn't mean this bottling corporation. [00:32:08] Speaker 00: All we meant was that there was substantial evidence to support that it would be complex, it would be difficult, it would be costly, yada da. [00:32:16] Speaker 00: Would that be sufficient? [00:32:18] Speaker 00: And in that case, why wouldn't we send it back to the board? [00:32:21] Speaker 03: Well, I don't think the court and deans have sent it back to the board. [00:32:23] Speaker 03: That was our first point. [00:32:25] Speaker 03: But if the hypothetical board decision that then were to go back and find, rely on evidence that they didn't cite on appendix 38 of why someone wouldn't have done that, the board didn't do that. [00:32:39] Speaker 03: And we think our evidence and the substantial evidence [00:32:42] Speaker 03: that supports that the person would know how to do this and that there's advantages of this shared cache so that a person wouldn't make this combination. [00:32:50] Speaker 03: But that's not something that's reflected in the board's decision. [00:32:52] Speaker 03: They never reached that point of the analysis. [00:32:55] Speaker 03: And if I may just remind that the board found that you would combine King and Arameli's teaching at least for the L1 and L2 caches. [00:33:02] Speaker 03: for they found the claims two and three would be obvious based on that combination. [00:33:07] Speaker 03: So I think it's a high hurdle to then say that, but you would stop at part of Erroll Milley's teaching, then you wouldn't be motivated to combine that teaching because of the changes that have to be made when there's a recognition that at least there would be an obviousness of combining King and Erroll Milley. [00:33:29] Speaker 03: I'm happy to respond. [00:33:30] Speaker 03: I realize I'm over time. [00:33:31] Speaker 03: I'm happy to respond on the claim destruction point, if the court wants. [00:33:37] Speaker 03: Two points, if I may, just quickly. [00:33:41] Speaker 03: In my friend's argument, he really was admitting that this is one limitation, plurality of bus segments. [00:33:48] Speaker 03: must comprise the plurality of flexible data channels. [00:33:51] Speaker 03: And you don't get those flexible data channels without switches. [00:33:54] Speaker 03: He pointed to appendix 1798 of the file history. [00:33:58] Speaker 03: I would point you to 1797 of the file history, where Pack explained that the multiple data paths are created by switches. [00:34:06] Speaker 03: And again, reply page 16 of our brief shows a diagram, no dispute about what's direct and indirect. [00:34:12] Speaker 03: You don't get the flexible data paths without indirect connections. [00:34:16] Speaker 01: Thank you, Your Honors, for your indulgence.